Copilot Studio | Vlad Talks Tech

AB-620 Exam: What to Expect, Difficulty, and Study Tips

Vlad Catrinescu — Wed, 19 Aug 2026 04:59:33 GMT

The AB-620 exam is the path to the brand-new Microsoft Certified: AI Agent Builder Associate certification, and after taking the beta a few hours ago, I want to share my honest first-look review. If you are deciding whether to sit for this exam, what to study, or whether you are even the right audience for it, this post covers everything I can share without breaking NDA.

Key takeaways

Passing the AB-620 exam earns you the Microsoft Certified: AI Agent Builder Associate certification.
Despite being tagged as an app maker exam on Microsoft Learn, the AB-620 is really built for developers and consultants integrating Copilot Studio with enterprise systems.
Microsoft scrapped the original AB-610 Copilot Studio Agent Maker exam and consolidated everything into the AB-620.
Expect heavy coverage of Azure AI Search, Microsoft Foundry, Microsoft Fabric, the agent-to-agent protocol, and Power Platform pipelines for ALM.
Testing agents and application lifecycle management makes up 20 to 25 percent of the exam, which is enough to make or break a pass.
If you are a Power Platform app maker, the Copilot Studio Applied Skills are likely a better fit than this exam.

What is the AB-620 exam?

The AB-620 is a Microsoft associate-level certification exam currently in beta. Pass it and you earn the Microsoft Certified: AI Agent Builder Associate credential. The official exam name is Designing and Building Integrated AI Solutions in Copilot Studio.

The exam tagline tells you most of what you need to know: Develop and integrate advanced, scalable AI agents using Microsoft Copilot Studio, Power Platform, and enterprise technologies to deliver robust solutions for organizations. In plain language, the AB-620 exam is about putting agents into production, not about the basics of agent building.

Why the AB-620 is really a Copilot Studio developer exam

Here is the most important thing to understand before you register: even though the exam is tagged as an app maker exam on Microsoft Learn, in my opinion that tagging is misleading. The AB-620 is built for developers and consultants who spend almost every workday building agentic solutions that integrate Copilot Studio with Microsoft Foundry, Azure AI Search, Microsoft Fabric, and more.

If you are a citizen developer who occasionally builds a topic in Copilot Studio, this exam will feel out of reach. If you live in Copilot Studio and routinely wire it up to enterprise services, you are exactly the audience Microsoft built the AB-620 for.

Having spent the last decade helping IT pros prepare for Microsoft certifications, I have seen plenty of exam mismatches between the listed persona and the actual content. The AB-620 is one of the more notable examples I have come across recently.

What happened to the AB-610?

A few months ago, Microsoft announced two Copilot Studio certifications: the AB-610 Copilot Studio Agent Maker and the AB-620 Copilot Studio Agent Developer. That two-track approach made sense. The Agent Maker exam would have covered the low-code persona, and the Agent Developer exam would have covered the integration-heavy persona.

Microsoft has since scrapped the AB-610 and consolidated everything into the AB-620. Instead of keeping the original Copilot Studio Agent Developer name, which would have been honest signaling, Microsoft renamed the exam to AI Agent Builder Associate. That broader name suggests both makers and developers belong here, but the content still skews heavily toward the developer side.

For the full backstory on these two exams, see my earlier post on the original AB-610 vs AB-620 Copilot Studio certifications.

What the AB-620 exam actually tests

The meat of the AB-620 exam is about production-grade agents and the advanced scenarios an enterprise will ask about before it ships an agent to production. Do not expect questions about turning generative AI on, what a topic is, or who should be authoring agents. That foundational content is not what this exam is measuring.

What you will see is heavy coverage of how to connect a Copilot Studio agent to enterprise services. Specifically, expect to be tested on:

Connecting a Copilot Studio agent to Azure AI Search
Connecting a Copilot Studio agent to a Microsoft Fabric data agent
Connecting a Copilot Studio agent to a Microsoft Foundry agent
Connecting to partner agents outside your tenant using the agent-to-agent protocol
Designing multi-agent solutions that span those systems

Real production agents today are rarely built entirely inside Copilot Studio. The AB-620 reflects that reality by testing your ability to design solutions that span multiple Microsoft AI services.

Testing agents and ALM with Power Platform pipelines

A second major area of the AB-620 exam covers testing agents and application lifecycle management. This bucket is marked at 20 to 25 percent of the exam, which sounds modest on paper but is more than enough to be the difference between pass and fail.

You need to know all the ways to test Copilot Studio agents and how to implement application lifecycle management for agents using Power Platform pipelines. The exam expects you to be able to take an agent from a development environment, through any required validation, and all the way to production while following a proper ALM process.

The takeaway is that the AB-620 does not assume you are only a builder. It assumes you own the full lifecycle of an agent, from build to release.

How to study for the AB-620 exam

The single best piece of advice I can give is to do every scenario in the exam objectives at least once before sitting down. Go to Microsoft Learn, pull up the AB-620 study guide, and work through the tasks hands on:

Connect a Copilot Studio agent to Azure AI Search.
Connect a Copilot Studio agent to a Microsoft Foundry agent.
Connect a Copilot Studio agent to another Copilot Studio agent.
Experiment with the agent-to-agent protocol so you know what it is and why it exists.
Build a deployment flow for an agent using Power Platform pipelines.

Why does hands-on practice matter so much? Because every associate-level Microsoft exam, including this one, leans heavily on order-list questions. You will be asked, “you want to do this, put the following steps in the correct order.” If you have not configured the scenario recently, those questions are punishing. If you have walked through it once, they become almost free.

The other point worth flagging is the question weighting. Microsoft did honor the study guide in the sense that nothing showed up in my beta that was not on the published exam objectives. However, the percentage of questions from each bucket felt off compared to the weights listed on the study guide. I cannot share which buckets were over-represented or under-represented due to NDA, but my recommendation is the same as it always is for a beta: study every skill area equally and do not lean on the weights as a planning tool.

Exam format, questions, and Microsoft Learn access

A few practical notes about the exam itself:

The number of questions is always between 40 and 60. I cannot share the exact count for the AB-620.
You have access to Microsoft Learn during the exam, just like every other associate-level exam. I did not end up using it, but it is there if you need to verify something.
Expect associate-level question types: scenario-based questions, order lists, and multi-step problems. This is not the difficulty level of a fundamentals exam.
Labs may show up. Microsoft does not publish whether a given exam includes hands-on labs, and I cannot share what was on my exam, but be prepared for the possibility.

AB-620 vs Copilot Studio Applied Skills

If you are a Power Platform app maker who does not spend all day in Copilot Studio, the AB-620 is probably too advanced for where you are right now. The better path is the Copilot Studio Applied Skills. There are two of them available, and they validate hands-on Copilot Studio skills without the integration-heavy enterprise scope of the AB-620.

If you want a deeper look at the lab-based credential for Copilot Studio, see my APL-7008 study guide for creating agents in Microsoft Copilot Studio.

The AB-620 is still a great goal if your job is to become a full-time Copilot Studio agent builder, or if you are already a developer integrating AI services across the Microsoft stack. For everyone else, the Applied Skills path will give you a recognized credential without the steep developer-focused prep.

What this means for you

The AB-620 is a high-value certification for a narrow audience: developers, consultants, and architects building production-grade agents that span Copilot Studio, Azure AI Search, Microsoft Foundry, Microsoft Fabric, and Power Platform. If that describes your day-to-day work, this exam validates the skills you are already using and is well worth the prep time.

If you are a citizen developer, business user, or IT pro who builds simple Copilot Studio agents on the side, this exam is not the right next step. Start with the Copilot Studio Applied Skills instead. They give you a recognized credential without asking you to master enterprise integration patterns you may never use.

Before you register, pull up the AB-620 exam objectives on Microsoft Learn and do an honest gap analysis. If you can do each task on the list without looking it up, you are ready. If half the objectives are unfamiliar, build hands-on time with the missing scenarios before scheduling the exam. Beta exams take longer to score, so plan accordingly if you are aiming for a specific certification deadline.

Prefer video? Watch the full walkthrough on my YouTube channel.

Ready to build your Copilot Studio skills the right way? I have over 100 courses on Pluralsight covering Microsoft 365, Power Platform, and Copilot, including hands-on agent building. Browse the full course catalog on Pluralsight to find your next step.

Copilot Studio Applied Skill APL-7008 Is Retiring June 30

Vlad Catrinescu — Thu, 25 Jun 2026 14:01:16 GMT

Microsoft is retiring the Create Agents in Microsoft Copilot Studio Applied Skill, known by its code APL-7008, on June 30, 2026. If you have been meaning to add this Copilot Studio Applied Skill to your transcript, you have until the end of June to earn it for free, and after that it is gone. Below is what the credential covered, what the retirement actually means for your transcript, and which Copilot Studio credentials you can still earn once it is gone.

Key takeaways

APL-7008, the Create Agents in Microsoft Copilot Studio Applied Skill, retires on June 30, 2026.
It is free, it is not proctored, and it runs as an interactive lab, so you can earn it in a single sitting.
Applied Skills never expire. Once you pass, the credential stays on your Microsoft transcript permanently.
After June 30, the Enhance Agents with Autonomous Capabilities Applied Skill and the AB-620 AI Agent Builder Associate certification are your remaining Copilot Studio options.
This is one piece of a much larger shake-up. We are not even halfway through 2026, and Microsoft has already retired or replaced 14 certifications and 13 Applied Skills.

What the Copilot Studio Applied Skill (APL-7008) covered

APL-7008 validated your ability to build a custom agent from start to finish in Copilot Studio. It was a well-rounded credential that touched almost every core part of agent building, not just one narrow feature.

The assessment covered creating the initial agent, working with topics and trigger phrases, and configuring the different nodes in a conversation. It also covered managing variables and entities, setting up the generative AI side with knowledge sources and actions, and pulling data from Dataverse to make the agent more useful.

As someone who earned this Applied Skill myself, I can tell you it was a genuinely practical credential. It mapped closely to the work you actually do when you build an agent for a real business scenario.

If you want the full background on the credential, I covered it in detail when this Applied Skill first launched.

Why retiring this credential is a surprise

I will be honest: this retirement caught me off guard. Copilot Studio is exactly where Microsoft is pushing companies right now, so I expected an update to this credential rather than a full retirement.

There is a reason behind it, though. Copilot Studio is going through a massive change as a product. The interface is changing, topics have been deprioritized and tucked away in the new experience, and the platform now leans on skills and more AI capabilities. A refresh would have made sense to me. A complete retirement, with less than two weeks’ notice, still surprised me.

This is not happening in isolation either. The pace of credential changes has been relentless this year, which I broke down in my roundup of every Microsoft certification change in 2026.

What an Applied Skill retirement actually means

Here is where Applied Skills work differently from regular certifications, and it is the part most people miss. Applied Skills do not expire. If you pass one, it shows up as an active credential for two years after the retirement date, and then it moves to your historical credentials.

More importantly, it stays on your Microsoft transcript forever as a credential you have passed. Retirement does not erase it. It simply means no new learners can earn it after the cutoff date.

That is why earning APL-7008 before June 30 is worth doing, even though it is going away. You keep the credential permanently, and right now, you can still take it for free. If you want the bigger picture on how these two credential types differ, I wrote a full comparison of Applied Skills versus certifications.

The Copilot Studio credentials you can still earn

Losing APL-7008 does not leave you without options. Two Copilot Studio credentials are available today, and they cover different audiences.

The first is the Enhance Agents with Autonomous Capabilities Applied Skill. It goes deeper and more niche than APL-7008 did, focusing on the autonomous side of agents, but it is still firmly in the Copilot Studio world, and it is still free.

The second is the AB-620 AI Agent Builder Associate certification. This is a brand new exam aimed at developers and advanced builders creating production-ready agent solutions. I took it myself last month, and I will warn you that it is not an easy one, certainly harder than the title might suggest. I compared it to the maker-focused option in my breakdown of the AB-610 and AB-620 Copilot Studio certifications.

What this means for you

If earning a Copilot Studio credential has been on your list, the next few days matter. APL-7008 is free, unproctored, and you do not need to schedule it. You simply open the assessment on Microsoft Learn and complete the interactive lab, usually in a single sitting. After June 30, 2026, that option disappears.

So the practical move is simple. If you want APL-7008, earn it before the deadline and keep it on your transcript for good. If you miss the window or want to go further, the Enhance Agents with Autonomous Capabilities Applied Skill and the AB-620 certification are both still available.

Whatever you decide, do not assume you can come back to APL-7008 later. Once it retires, there is no way to earn it again.

Ready to earn APL-7008 before it retires? You can take the Create Agents in Microsoft Copilot Studio Applied Skill for free on Microsoft Learn until June 30, 2026. Start the assessment on Microsoft Learn. If you want to sharpen your agent-building skills first, my Pluralsight courses walk through Copilot Studio step by step.

Prefer video? Watch the full walkthrough on my YouTube channel.

Copilot Studio Certifications Are Coming: AB-610 vs AB-620 (What Microsoft Shared So Far)

Vlad Catrinescu — Mon, 19 Jan 2026 12:52:00 GMT

Microsoft is working on two new Copilot Studio certifications, and they target very different audiences. In this video, I break down the public info available today, so you can understand what’s coming and how to prepare.

✅ Copilot Studio Agent Maker (AB-610): Built for makers, business users, and IT pros creating low-code agents in Copilot Studio.

✅ Copilot Studio Agent Developer (AB-620): Built for developers and advanced makers building enterprise-grade agent solutions, including integrations and advanced capabilities.

I walk through what Microsoft has shared so far, including the target audience and the skills measured for each exam!

Want to learn from some of the top professionals around the world? Join me at the M365 Conference and get $150 off with the discount code VLAD150!

Watch my 100+ courses on Pluralsight

Video Summary

Two new Copilot Studio certifications are coming: AB‑610 Agent Maker and AB‑620 Agent Developer. They’re not live yet, but Microsoft is far enough along that we’ll likely see them in the next 2–3 months.
Agent Maker (AB‑610) targets business users, subject‑matter experts, and low‑code builders who want to create agents using Copilot Studio’s full experience. Expect skills around flows, topics, grounding with data, and best practices — though some items (like autonomous agents and parent/child orchestration) feel a bit advanced for “makers.”
Agent Developer (AB‑620) is for professional developers and advanced makers building enterprise-grade solutions. It dives into Microsoft Foundry, MCP, custom connectors, API integrations, Fabric data, computer use, multi‑agent systems, and advanced ALM — the real “pro dev” track.
Key difference between the two: the Maker exam sticks to common data sources (SharePoint, OneDrive, websites, Dataverse), while the Developer exam jumps into enterprise systems like SAP, ServiceNow, Azure AI Search, and cross‑platform model integrations. Developer skills also require GenAI knowledge, integration patterns, and ALM.
Timeline looks soon: Microsoft has already started recruiting item writers, with alpha reviews slated for early February 2026 — a strong sign the exams are close to release. In the meantime, you can already earn Copilot Studio Applied Skills today.

For more information, read the transcript blog below, or watch the video above!

Transcript

It’s no secret that 2026 will be all about AI agents, and with Copilot Studio being the primary tool for building AI in 2026, many people are asking: Is there a certification for it? Well, I have good news for you because soon the answer will be yes. And if you care about staying relevant and showcasing your agent‑making skills, this video is for you. Microsoft recently made public that they are working on not one but two different certifications for Copilot Studio, and they are for different audiences. The first one is a Copilot Studio Agent Maker certification, and the second one is a Copilot Studio Agent Developer certification. Now, those are not fully announced yet, so they’re not in beta, you cannot take them today, and a few things might still change, but as you’ll see throughout the video, we are really far along. So, in this video, we’re going to cover everything you need to know about them: the skills measured, the detailed target audience, and really everything you need to know to start getting ready for them.

Let’s get started by covering the Copilot Studio Agent Maker certification, which should be exam number AB‑610. Now, I have to say, just from the name of it, I’m really a big fan of it because this reminds me of the Power Platform App Maker certifications. For those of you who remember—rest in peace, exam number PL‑100—I still miss you, and I think it should have never been cancelled, but it’s great to see Microsoft focusing on makers again.

Let’s start by covering the target audience. Microsoft says this certification is targeted towards makers who build low‑code custom agents with the full experience of Copilot Studio. They typically work as a subject matter expert, business user, and even an IT professional focused on creating agentic AI solutions for organizations. Again, this seems to be aimed at your makers and not at consultants or developers who really build those solutions all day long.

Now, as a candidate for this certification, Microsoft says that you need to be able to understand agent requirements, create and define agents, create agent flows, test and validate agents, implement agent best practices, and integrate agents with data, and, of course, build instructions. Now, a few highlights in here: Create agent flows, so we will need some Microsoft Power Automate knowledge, and it’s really important when you create a solution, so I’m happy to see that in there. And we also have, you know, implement best practices, which is really, really vague for an exam, but we’ll cover more details in just a minute.

And now, as an agent maker, you need to be able to create agents or experiences that are conversational and autonomous, they are grounded with data, perform actions, and connect to Microsoft 365 Copilot. I will say that this so far makes sense. The only thing that I’m a bit surprised about is the autonomous agents part. I feel like a maker would not really create an autonomous agent, especially with the recent security breaches that we have seen around that. So I really see that autonomous part more towards a developer or functional consultant role, as Microsoft puts it, not an agent maker, but that’s what Microsoft has there so far.

If we look at the names of the official skills measured so far, Microsoft has split the certification into three main buckets: Create a Copilot Studio agent, use topics and tools in Copilot Studio, and deploy and monitor agents. But I told you we’re pretty far along; we can even go deeper into detail. So, to do that, I opened up my browser over here, and I’m getting this information because Microsoft and the amazing Liberty Munson put out a blueprinting survey opportunity. And this is all public information. This is where Microsoft gets the community’s opinion on what the most important things are that should be covered in the exam, the weighting for each one of them, and so on.

But the really cool thing about when we get those is that we can extract a lot of information. We covered a lot of it here already in the slides, but a few things that I want to add before we go into the skills measured: You see here they have “you collaborate with,” and then you have the M365 admin, Power Platform admin, Copilot admin, Copilot Studio Agent Developer—the other certification we’ll talk about in a few minutes—Copilot Studio admin, Agentic AI Business Solutions Architect, which is exam number AB‑100 (this just went out of beta today as I’m recording this, so let’s say start of January for those of you watching it later), and Copilot Studio Architect.

Now, why I’m saying this is interesting is because, you know, this is already a certification, this is a certification, this is a certification. Are we going to get certifications for the rest of them? I would love a Power Platform admin certification. I think that would be awesome. A Copilot admin certification—that would be amazing. So maybe those are roles that Microsoft defined, and we might get certifications for them in the future. I would absolutely love that.

But for now, let’s go back to our Agent Maker certification. What I want you to focus on here now is really the side on the left. So the way that we have it split, we have our big skill measured—the one that has the percentage—create agents in Copilot Studio. Then we have sub‑skills: design Copilot Studio agents, create an agent, and configure agent capabilities. And for each one of them, Microsoft goes really deep into the details.

So, for design Copilot Studio agents, you need to be able to recommend Copilot Studio agent components, design autonomous agents, differentiate the types of agents, including declarative, managed, and custom, and design parent‑child agent orchestration in Copilot Studio. Now I’ll be honest: For some of the,m I really agree with those, but I feel like makers creating parent‑child agent orchestration or autonomous agents—for me, that goes beyond the agent maker role and goes more into that consultant or developer role, but that’s just my opinion on it.

Then let’s go over to create an agent. You need to be able to create an agent using a template, create and configure an agent manually, create a declarative agent, and enhance an agent created from the light experience. Now, this is really funny because Microsoft renamed the agent builder to Copilot Studio Light, which is what is referenced here in September/October 2025, but they already renamed it from that. You know, Copilot—Microsoft marketing can’t even keep up with their own renames. It’s now called the agent builder experience in Microsoft 365 Copilot. Wow, what a name. I’m sure by the time the exam is live, this will get updated, but it’s still funny to see.

Then, for our final pillar, we have configured agent capabilities: Modify agent purpose and instruction—this makes sense. Configure knowledge sources, including uploaded data, SharePoint, OneDrive, websites, and Dataverse. Now I’ll pause here a second because right now it seems that this certification is only focused on your most common data sources: uploaded data, SharePoint, OneDrive, public websites, and Dataverse. This will be really important when we talk about the developer certification soon and the differences with that one.

But let’s move on. Add suggested prompts, configure agents to work autonomously, configure generative AI settings, and configure an agent to use child agents. Now look, I think most of this is awesome except the autonomous and configure agent to use child agents. I think for me that’s beyond the agent maker’s scope. And by the way, if you agree with me or disagree, please let me know in the comments below. And if you find videos like this useful and you want to keep up with Microsoft credentials, make sure you like the video and subscribe to the channel so you’re the first to know about them.

Okay, let’s move over to the next page now, over here and let’s take a look at the next skills: use topics and tools in Copilot Studio. Well, first of all, create and manage topics. Here, the sub‑skill: evaluate and update system topics—really important. Create custom topics, add business logic to topics, so your conditions, things like that. Add agent flows to topics, configure agent responses, and add tools to topics. So not only agent flows, but you might want to call a prompt in the AI builder as an example.

Configure advanced agent responses with custom prompts—well, this is the one that I just mentioned, but you really need to know all the different tools. I’m surprised they don’t have—okay, here it is: add tools to an agent, add agent flows to an agent, add connectors to an agent, which is the one that I was finding missing. Create and add custom prompts to an agent. So, honestly, a bit more of a repetition here. I guess you add them directly in the topic, while this, you add them at the agent level, and this can of course be very different, whether you have orchestration on or off, which I’m sure will be a big part of the exam.

But now, create and monitor agent flows. So may include, but is not limited to: create an agent flow, configure actions and connectors, monitor agent flows, and describe input and output parameters. So this is a lot of Power Automate, let’s be honest. And we all know that the real advantage of an agent over a chatbot is that agents are able to do things. How can they do things? Using Power Automate agent flows. So being proficient in Power Automate will be really important for this certification, as well as the few connectors that we have in agent flows that we do not have in Power Automate, like the advanced approvals, for example. I’m sure it will be a big part of this exam.

Now, moving on to the last one: deploy and monitor agents. So: deploy to Microsoft channels and the demo website. This is really interesting because to deploy it to the demo website, you actually need to have an anonymous agent, which, in my opinion, goes beyond the role of this certification. It’s not what a maker should do. Makers create agents for M365 Copilot chat and Teams, things like that, not necessarily the public one for the demo website.

But anyway, manage connections, share an agent, and identify different channel deployment requirements, including authentication. So a lot of stuff to know, a lot of it makes sense. I’m not sure that deploying to non‑Microsoft channels should be there, but it’s in there right now. Now, monitoring agents: evaluate agent performance and configure agent savings. I love that they put the agent savings in there. It’s just a small feature. Basically, for every time the agent answers a question, you can put a dollar amount, and at the end of the month, you have a nice big number to show to your boss on how much your company saved. It’s something Microsoft is really pushing, so I’m not surprised they added it in the exam, but it’s such a small feature.

Finally, manage Copilot Studio agents to environments: create a solution, export and import agents by using solutions, and add existing agents to a solution. Now, I’ll be honest: for me, I don’t think I’m saying that that is important. Don’t get me wrong—I think consultants should know that, developers should know that—but I’m not sure that your average agent maker creates stuff that gets deployed into solutions or that needs ALM and things like that. So I’m not sure how much agent makers need to know that or have access to creating solutions and things like that. For me, this is a bit beyond what an agent maker should know and goes more into that consultant or developer role, but Microsoft has it in here for now. So I would like to have this at the “design” or “describe” verb instead of “create, export, and manage,” but this is what Microsoft has for now.

So please again, let me know what you think about this certification. So far, the Agent Maker—other than a few things—I really love it. But with the Agent Maker covered, it’s time to switch over to the Copilot Studio Agent Developer certification, which should be exam number AB‑620. This certification is targeted towards professional developers and advanced makers who build, extend, and integrate custom agents for enterprise‑grade solutions. They typically work as IT application developers, consultants, or ISV partners focused on creating scalable AI solutions for organizations or customers. So, as you can see, this is really that next‑level certification for developers that all they do is build agents all day long—that is their day job.

As a candidate for this certification, you’re able to integrate agents with Microsoft Foundry. So we’re going even beyond Copilot Studio—we’re talking Microsoft Foundry here, which is interesting. You’re also able to integrate agents with model context protocol servers or MCP, integrate agents with custom connectors, integrate agents with APIs, integrate agents with Microsoft Fabric, automate tasks with computer use, and integrate agents with connectors. So wow, as you can see, we’re talking custom connectors, we’re talking computer use in Copilot Studio, we’re really taking our knowledge to the next level.

As a Copilot Studio Agent Developer, you’re also in charge of creating multi‑agent solutions. Again, this is where I think that a multi‑agent solution should fit, not in the maker one. You need to be able to create agents with enterprise knowledge sources like ServiceNow, SAP, and more. And this is where I was saying before that remember, the Agent Maker seemed to only focus on SharePoint, OneDrive, public websites, Dataverse, things like that, where the developer—we’re talking third‑party tools. So again, we’re taking our knowledge to the next level.

You also create advanced agent topics and tools, computer use agents, and agents that perform advanced actions via APIs. So, for me, this fits perfectly with the Copilot Studio Agent Developer role. If we go to the skills measured and we talk again about what the three buckets for this one are as well: we have plan and configure agent solutions, integrate and extend Copilot Studio agents, as well as test and manage agents.

So let’s go back in the browser and check out the detailed list for this exam as well. So here we’re back on a very similar page, but for a blueprinting survey for the Agent Developer certification. A few things—again, we covered a lot of this already. A few things that I want to show you: new role here, Microsoft Foundry Administrator, which is interesting—and a few things you should be familiar with: Power Fx, Dataverse, Power Platform environments and components, Microsoft Foundry, adaptive cards—all things that for me make sense as a Copilot Studio dev.

Now, for the prerequisites we have—wow, look at this—this is so much more advanced: intermediate knowledge of generative AI concepts such as models, orchestration, retrieval‑augmented generation, model context protocol, agent‑to‑agent protocol, and more. So as you can see, you really need to have that GenAI concept knowledge, not only Copilot Studio. Prompt engineering experience, of course, experience with REST APIs and integration patterns,and familiarity with core Microsoft cloud platforms.

This is the first time that we see that exam number for the Copilot Studio Agent Maker—Copilot Studio Agent Maker AB‑610 knowledge and skills. Now, this is really important that they put this inside the prerequisites because I do wonder: Will Microsoft make the Copilot Studio Agent Maker a prerequisite for the Agent Developer, or do they just put it here as a recommendation, like “Hey, make sure you have the Agent Maker skills before you go to become a dev”? Or will they make it official, where you actually need to pass the Agent Maker before you can get the Developer certification? So that will remain to be seen.

So let’s go and now take a look at the details here. Plan and configure agent solutions: the sub‑skills inside: plan an agent solution, create and monitor agent flows, configure topics—and those are the three. Now may include but is not limited to: plan integrations—I’ll zoom in even a bit more—plan integration with enterprise systems, plan identity strategy, plan channels and deployment, plan responsible AI strategy, evaluate security and governance considerations—really important—plan reusable agent components, and design agents for internal and external audiences. You see, this is really at the “plan and design” verb, but a lot of important stuff is in here.

Now, create an agent for creating and monitoring agent flows: create an agent flow, create a human‑in‑the‑loop agent flow. Now, not sure how I feel about this name, but that’s how Microsoft names them. Things such as the advanced approvals—they are in the human‑in‑the‑loop connector category, so you need to know those. Configure actions and connectors, monitor agent flows, describe input and output parameters, and implement error handling in agent flows.

For me, I’ll be honest, I feel like the Agent Maker one and Developer one should be the same because even a maker needs to know how to implement error handling. So they only added a few extra things in here, but yeah, I feel like, you know, error handling could have been in the Agent Maker as well.

Now configure topics: add agent flows to a topic, configure agent response formatting, add tools to a topic, configure agent advanced responses with custom prompts, configure agent responses with custom knowledge sources—so again, we’re going beyond those built‑in knowledge sources; we’re going to custom knowledge sources for this certification. Configure advanced agent responses with API and send HTTP requests, configure the generative answers node, configure adaptive cards, and manage variables. We didn’t have managed variables in the previous certification, the Agent Maker, but I definitely think you need to know how to do it now.

Okay, let’s move over to the next page. Next, we haveto integrate and extend Copilot Studio agents: connect to enterprise knowledge sources, connect to Copilot connectors—which by the way is the new name for Graph connectors—connect to Power Platform connectors, and connect to Azure AI Search. Again, we’re moving beyond just M365; we’re moving to systems outside the Microsoft ecosystem.

Even add tools to an agent, configure and monitor computer use for an agent, configure model context protocol tools, add a tool using an existing custom connector, and add REST APIs to an agent. So again, I think those are all awesome. Configure multi‑agent collaboration from Copilot Studio—this is where it gets interesting. Design multi‑agent solutions in Copilot Studio and integrate a Microsoft Foundry agent. I’m surprised that for this certification we’re going to custom agents in Microsoft Foundry. I do think it’s important; I do wonder if it’s different roles, but having that knowledge and knowing when to pick Foundry versus Copilot Studio—definitely important.

Integrate an existing Copilot Studio agent, integrate to Microsoft Fabric data agent—so really interesting. You need to have some Fabric knowledge as well. Create a multi‑agent solution by using the agent‑to‑agent protocol. So this is why they said before you need to know that protocol—you need to know what it is and how to use it.

Now, integrate agents with Azure: configure generative answers using Azure AI Search with Microsoft Foundry—again, this is very useful for your knowledge outside of Microsoft 365. Configure custom prompts to use the Microsoft Foundry model catalogue—so what if you want to use Groq, for example? What if you want to use something outside OpenAI or Anthropic? You have to go to the Foundry. So this is interesting. And monitor agents with Application Insights—really interesting. I love it.

Finally, for test and manage agents, you need to evaluate agent performance, you need to be able to create a test set, choose an evaluation method, and review test results. So this is something that Microsoft announced at the Power Platform Community Conference 2025, so I love that this is in here. I would have loved to see some of them in the Maker exam as well, but I love that it’s in the certification—definitely important.

And implement application lifecycle management for Copilot Studio agents—again, very important for developers. Create a solution, add existing agents to a solution, create and use environment variables, and implement and extend Power Platform pipelines. Remember when I said for the Maker one, they shouldn’t have to deal with solutions other than maybe be aware of them, but they shouldn’t handle them? This is where, for me at least, this topic should be, because developers manage solutions—makers do not.

But this is it for the skills measured inside the Copilot Studio Developer exam. Now with the details covered, I really want to hear from you in the comments below. What do you think about those two new certifications? Do you think they are interesting for what people need today, or do you think Copilot Studio is still marketing‑focused and most enterprises are not using it yet? And if you care about them, what do you think about the skills covered? I shared with you my opinion; I want to hear back from you.

But my personal take is that I love the Copilot Studio Agent Maker. I’m going to take that one. But since I’m not a dev, I don’t think I’ll go for the Copilot Studio Agent Developer. I feel like, you know, I can do about 80% of it, but there’s definitely a 20% which I never touched. But never say never.

Some of you are probably really excited, and you’re wondering when those are coming out. Honestly, I think it’s sooner rather than later, because just a few hours before recording this video, there was a LinkedIn post from a vendor which talked about the exams and was already recruiting to write the questions. And from what I see in that post, it seems that the review for at least the alpha version of the questions will be done on February 2nd, 2026.

So, of course, there’s a lot of work going on between the alpha version and the actual exam going live—you know, there’s a lot of work in there—but Microsoft is so far along. I’m like 95% sure those exams will go live. Well, less sure, but still pretty sure we’ll see them in the next two or three months. Maybe Microsoft Build—I don’t think we have the dates for that yet; it’s usually May—but I would really like to see them, you know, in March, April timeframe.

But I even have some even more great news for you, because if you want to get a Copilot Studio credential today, Microsoft Copilot Studio Applied Skills are available today. So you don’t have to wait. And on the screen right now, you actually have a link to one of my Pluralsight courses, which covers how to get started creating a Copilot Studio agent, as well as a video where you can learn more about the APL, the “Create a Microsoft Copilot Studio Agent.” I hope I’ll see you in that video, and I’m looking forward to reading your comments on those two new certifications.

AB-900, AB-100, AB-730, AB-731: Making Sense of Microsoft’s New Copilot Certifications (with Tiago Costa)

Vlad Catrinescu — Tue, 30 Dec 2025 13:33:00 GMT

Microsoft just reorganized its certification lineup into a new AI Business Solutions track and introduced four new Copilot/AI certifications. If you’re wondering what these exams actually cover, who they’re for, and which one makes sense for your role, this video will save you a ton of time.

In this discussion, I’m joined by Tiago Costa (Microsoft MVP & MCT), and we break down what’s new, what surprised us, and what to watch out for as these certifications move from beta to GA.

In this video, we cover:
✅ What changed with Microsoft’s certification structure (MB/PL/MS → AI Business Solutions)
✅ The 4 new certs: AB-900, AB-100, AB-730, AB-731 — who each one is really for
✅ AB-900: wide objective domain, what “fundamentals” actually means, and how to prep
✅ AB-100: why it’s tough, what kind of role it targets, and why you should actually study for it
✅ AB-730 vs AB-731: business user vs transformation leader, and the mindset shift required
✅ What Microsoft Applied Skills are good practice (and why they’re underrated)

If you’re planning your 2026 certification goals around Copilot and AI, this one’s for you.

Want to learn from some of the top professionals around the world? Join me at the Microsoft 365 Conference and get $150 off with the discount code VLAD150!

Watch my 100+ courses on Pluralsight

Watch my playlist on Creating Agents for Microsoft 365 Copilot!

Video Summary

Microsoft reshuffled the certification world, merging MB, PL, and MS into the new AI Business Solutions track — and introduced four fresh AI/M365 Copilot certifications that are already shaking things up.
AB-900 (Copilot Admin Fundamentals) is fair but massive — the objective domain is so wide that even seasoned M365 pros don’t touch everything it covers daily. Great for beginners, but instructors will struggle to teach it in the “recommended” 5.5 hours.
AB-100 (Architect) is no joke — extremely broad, pulling in M365, Power Platform, Dynamics 365, and even Foundry. Tiago called it “very challenging” and recommends serious prep before attempting it.
The business-user certs (AB‑730 & AB‑731) surprised everyone — 730 goes much deeper than expected, requiring you to know exact steps, tab names, and agent-building workflows. And 731 flips your brain into an executive mindset — ethics, governance, responsible AI, and organizational readiness.
Applied Skills are underrated weapons — they’re free, hands-on, and perfect warm‑ups for several certs. Both of us recommend taking the Copilot‑related Applied Skills before jumping into the business-user exams.

For more information, read the transcript blog below, or watch the video above!

Transcript

In the past few months, Microsoft has completely changed the way they organize certifications. By merging the MB, PL, and MS series into one new bucket called AI Business Solutions, we also got four new certifications related to AI and M365 Copilot.

Today, I have the pleasure of being joined by Tiago Costa—old friend, Microsoft MVP, MCT, and you probably have all the titles at Microsoft, to talk about those. Tiago, you’re super popular in the learning community, but for the few people who don’t know you, can you introduce yourself?

Yeah, sure. Basically, I work as an AI and cloud architect, and I also teach. I teach a lot in the classroom. It’s one of my big passions—being in the classroom and passing knowledge to new generations, and even current, even older generations too. The idea is always to pass new knowledge to people. And, you know, like Vlad, I like to do conferences and speaking. So that’s kind of me—focusing a lot on the AI and Azure space.

Okay, that’s amazing. And between us two, we actually took all the AB certifications. Yeah. So first of all, I want to get your opinion. So far, we have four of them. We have an architect-level one, AB‑100. AB‑100. We have a Copilot Administration Fundamentals, which is a 900‑level course. Yes. And we have, for the first time, two of them for business users. Was that on your bingo card this year—that we’re going to get business user certs?

No, definitely not. I was hoping to have some more Copilot fundamentals. I thought, “That’s definitely needed,” and I think that wasn’t a surprise when they launched that one—at least not for us in the industry. But not the architect, and definitely not the business users. That one, I was like, “Wow, that could be a good idea,” because business users can actually do a lot of push for Copilot and AI in companies and really make use of the AI tools.

It’s the key to unlocking significant productivity. It’s not just having the tools—it’s knowing how to use them and the power behind them. So I think it’s a great move. Okay. And you mentioned Copilot Fundamentals, but in a way, I see the 730—the AI Business Professional—as Copilot Fundamentals. And I know you’re yet to take that one. Yes, I have it scheduled. It’s a lot about using Copilot, of course, based on the name, and creating agents. Creating agents, yes. We’re going to have a lot of fun with that one. But let’s start with the one we both took, the AB‑900. Yes.

What did you think about that one for starters? Well, I think it was a very fair certification. We took it in beta, so—It might change by the time… Exactly. For the audience, you need to take our conversation with a grain of salt because things might change when the certification is GA—generally available. But in beta, I think it was a good, fair exam. I’m not going to lie—there were two or three questions that were a bit weird. Oh, that’s always the case. One of them, I thought, “Well, this needs some polishing,” but that’s why it’s a beta. I gave my feedback—“Hey, this should be more like this”—so the exam authors can polish things. And they do. But overall, it was a fair representation of a fundamentals certification.

For people who don’t know what a fundamentals certification is, it’s for people getting started in that technology. It’s not for people with five, six, seven, or eight years in the industry. For those people, fundamentals should be really easy. Should be, yes. Should be, at least. One thing I noticed—and I want your opinion on it—I fully agree with you that the questions matched the objective domain. But the objective domain is so wide. Yes.

I feel like in most enterprises, your average admin doesn’t touch everything from Exchange Online to SharePoint to Purview to SharePoint Advanced Management to Entra ID. There are just so many things. And again, this is a fundamentals exam. You don’t need to know how to do things—you need to know what things are and where they are. You’re not going to get “Write the PowerShell cmdlet.” Definitely not. Fundamentals are usually at the “describe” level: describe what X is, what Y is. Still, as you said, the objective domain is really wide. There’s a lot of stuff.

For example, Purview—I don’t work with Purview daily. I know what it is, I know people who do it 100% of the time. I interact with them, so for me, this certification was fine. But I guess a lot of people taking this fundamentals certification—the exact audience—they might not have experience even at a “describe” level. But I also see this as a challenge and an opportunity. It exposes them to things outside what they normally do. And since it’s “describe,” they don’t need to go deep.

Okay. You’re saying something interesting—that it motivates people to learn other things. At Ignite, I talked to someone who said something that stuck with me: “This is what Microsoft wants an M365 admin to be.” They want admins to know Purview—not configure it, but at least know it exists. So if a business user asks, “How do I protect this content?” they don’t say, “I have no idea.” Exactly. It’s knowing that those features are there. Another question: you’re probably going to teach classes on this. Very probably. Or create my own training—definitely creating my own. You also have Learn with Tiago, with great on‑demand training. Yes. We don’t usually compete—you have the Azure stuff, I have SharePoint—everybody’s happy. But I find it will be tough to balance how to teach this and still keep it in Microsoft’s recommended 5.5 hours.

You and I like to go beyond slides—we like demos. But now we have to show SC‑900, MS‑900, Copilot… it’ll be a challenge. Definitely. I haven’t started planning content yet, but it’s on the list. When I saw the objective domain, I had the same thought: “This will take 12 to 15 hours to teach properly.” But I can’t do that for a fundamentals course. I’ll need to trim things. It’ll be hard for you too—deciding what to demo or skip. Exactly. Problem for 2026: Tiago and Vlad. Exactly. Even SharePoint Advanced Management is one line—but it took me an hour and a half to present it. Exactly. Do I just explain what it is? Do I demo it? What do I choose? Problem for future us. Definitely.

Next up: the architect exam—AB‑100. Yes. This is the one I didn’t take. Looking at the objectives… Microsoft didn’t even add AB‑100 to the certification poster. They forgot. But they have the objective domain available. It’s an exam where, if you think AB‑900 is wide, AB‑100 is even bigger. That’s why I didn’t take it. To be honest with the audience, it’s been a big challenge. I work with that stuff—that’s fine—but there are touches of things I don’t do daily. Especially Dynamics. There are a lot of Dynamics integrations. And Copilot agents in Dynamics that you need to know. So I think this exam is for a very specific audience.

A lot of the things—I look and think, “I have no idea.” Like: “Design connectors for Copilot in Dynamics 365 Sales.” I’m lost on two levels. You’re brave for taking it. Yeah, I took it. I have to confess—life got in the way of studying. Zero prep. I just thought, “Let me check the exam.” I still don’t know the score because it’s a beta. You only know weeks later. But I found it very, very challenging. I bet. Yes. Some questions, I thought, “Oh yeah, definitely this one.” Probably half of them. The other half—some you can figure out, but others are totally out of my knowledge because I don’t work with that stuff. So if you want to take AB‑100, my advice is: prep properly. Videos, classroom training, Microsoft Learn—read everything. At the time of recording (December 2025), there’s nothing yet.

Maybe by early 2026, they’ll release something. But prep very well. Maybe I’ll take it in the future. I’m laughing at how much they’ll need to rename. Yes. All of them. AI Foundry—Azure Foundry—now Microsoft Foundry. Dynamics 365 Copilot for Sales—that changed. Now we have Work IQ. There’s nothing about Work IQ because these certs were released before Ignite. Exactly. Agent 365 wasn’t there yet. Now it is. People will get a bunch of updates. That certification needs major updates before GA—wording, product names, and new products like Work IQ.

Final question for AB‑100: In your experience, does a role like this exist in enterprises? Someone who does Power Platform, Dynamics, Foundry—all of it? Or is this role a unicorn? Depends on the company’s maturity and the market. In North America, maturity is higher. In Europe, not so much. We’re recording this in Europe right now. I don’t see many companies here with someone who does all this. Some do—big, mature ones. But most?

No. They’ll have multiple consultants—PL‑2000s, PL‑400s, Power Platform architects. But not someone who does everything. Again, I wonder if Microsoft put this out to push the market. Yes. And I love seeing Microsoft lead, not follow. With the products and plans, this role will be needed. Companies need this. It’s good that Microsoft is pushing it.

Next up: AB‑730, AI Business Professional. You didn’t take this one yet. Not yet. I’m taking both business-user certs next week—Thursday and Friday. You’re going to laugh, but I took this at Ignite, and I found the 730 tougher than the 900. Oh really? Yes. Two reasons: AB‑900 is more of what I do daily. And I studied. This one—“easy business user”—so I thought I didn’t need to study. I should have studied. It’s not wide, but it’s deep and specific. Lots of: what tab do you click, what section, what tier supports what. Things I do on autopilot—I don’t remember the names. So I should have reviewed. So the audience gets this: even Vlad needs to study. 100% of the time. Don’t feel too confident going into exams. Even if it’s your daily job, there are things you don’t do often.

Good advice—because I wasn’t planning to study. Now I think I need to review. This one isn’t that many things. Within an hour, you could build everything hands-on. But you need to pay attention to every click—the name you type—everything. I cannot wait to make training for this one. It will be fun, but very different. Business users in a Microsoft certification class—it’s a different dynamic. Yeah. Business users come with a problem they want solved. You don’t know their company, their constraints, or their policies. It’s challenging. This exam requires: “You need to learn this. You need to remember this.” Not: “How do I solve my problem?” Exactly. But I’m excited to create content.

Now the last one—731. The AI Transformation Leader. Why do the numbers feel random? 731, not 740. It feels like the lottery. This one was very interesting. You’re probably more suited for it than I because it has both M365 Copilot and a lot of Azure. Yes, I already saw that. You need to know M365 Copilot, licensing, Foundry, Document Intelligence, Speech Services, Vision, Search—everything. And licensing differences. You’ll have no problem with that. But the part I also noticed—and I actually studied for this one—there’s a lot of adoption specialist content. And ethics. The Microsoft AI ethics framework. I do adoption in training, but I’m not usually the one saying, “First step: hire a Chief Ethics Officer.” Does it make sense for a transformation leader?

Absolutely. So I agree with the role. But for IT pros taking this exam, you must think like an executive. Learn Microsoft’s approach to ethics and responsible AI. It’s very different from the technical view. These topics are often covered in executive tracks at conferences—not tech tracks. Exactly. Executives don’t care what service does what. They have people for that. They care about governance, privacy, and ethics. High-level concerns. This certification touches fairness, responsibility—you can intentionally or accidentally introduce bias in a model. You need guardrails. This exam talks about that. It’s very good. They need to update wording—Work IQ, Agent 365, Foundry—but those are easy fixes before GA. For IT pros: review Microsoft Learn, understand how Microsoft pitches AI to executives. You need to think like an executive. Exactly.

Now, a question for you: should IT pros take this certification? Great question. Do they need to? No. Should they? I think consultants should. It forces you to understand how executives think, what their concerns are, and how to present to them. IT pros think about implementation. Executives think about governance. So it’s recommended. But I wouldn’t force it—no mandatory KPIs, no bonus requirements. But it’s a useful perspective. I agree. It’s always good to understand leadership thinking. And Microsoft may force this anyway for partner designations. Yes. So it might not be optional later. Exactly. And you’re always more valuable with certifications—especially in consulting. We covered all of them.

One final question: applied skills. Have you taken any yet? Yeah, I think I have two from Copilot, around thirty total. My goal is to have all of them. Catch them all. Yeah, kind of. Even if I don’t work with that stuff, I want to know what’s out there. It helps prove I know the basics. And pushes me to learn new things. I started my career as an IT pro on Unix and Solaris. Then I became a dev. I reinvented myself several times. I worked with cloud before cloud was cool. Now AI. It’s important to keep learning. You might find something you love.

For the Copilot applied skills I’ve done, some without prep—I struggled but passed with 100%. But I had to search for where stuff was. So read the Learn modules—spend 3–5 hours. Find content online—videos, etc. Vlad also has amazing content. I also have some content. Do you recommend applied skills as practice for certifications? Depends. Some align well, some don’t. But for some exams: absolutely yes. For example, for PL‑2000 or M365 Copilot, applied skills. Doing those first is good practice. They’re free. Not proctored. You can do them anytime. Just need internet. Exactly. You press a button, and you get a virtual machine with goals—not steps. Achieve the goals. They’re nice. If you fail, you can take them again after 72 hours. I had one crash on me, so it was marked as failed. It happens. You can take it again. I’ve failed applied skills before, too. Exams are hard even for us. Especially for new certifications when content doesn’t exist. Exactly.

Well, Tiago, thank you so much. I’m not going to say good luck—I’ll say break a leg—on the certifications next week: 731 and 730. It would be awesome to connect after and hear what you think. Hopefully, my tip to study helps. It can’t hurt. Now that you say that, I was planning not to study. I’m questioning myself. Maybe one hour of proper hands‑on and I’ll be good. One last question: if Microsoft listened to you with no questions asked, what are three certifications you want them to create in 2026? Ooh, very good. I’d love to have that power. I’d say: first, a pure dev certification—coding. Even with “vibe coding,” coding is more important than ever. Something around .NET, Python—core languages. Maybe easier to have Python now with AI.

Second: a pure dev certification for SharePoint. How to customize SharePoint—not the other stuff. I’m surprised you said SharePoint. I think it’s a miss. We used to have MS‑600—rest in peace. Rest in peace. But great for video views when they retire certs. True, but I’d rather have the certs. Third: an advanced Azure administration cert. More advanced than AZ‑104. Not an expert-level cert—but a second-level admin cert. That’s interesting. I think a collaboration admin cert—SharePoint, Teams, Planner, Stream, To Do, but not Teams Voice. Only collaboration. That would be awesome. I wonder if we’ll get an associate Copilot administrator cert—like AB‑900, but associate level. That makes sense. I think they might do that. My third: bring back PL‑100—Power Platform App Maker. The gap between PL‑900 and PL‑2000 is too big. And with agents and vibe coding, makers matter. There’s a need for that. Let’s hope. Let’s watch this video in a year and see if Microsoft listened.

Yes. We need to record this again in a few months as well. We don’t know where ESPC is next year—we’ll find out tomorrow. Wherever it is, we’ll record a follow‑up and see if we were right. But right now, the lights are closing behind us, so I think it’s time to finish the video. Tiago, thank you so much. And for everybody watching, if you want to connect with Tiago, you’ll have all his socials in the description below, as well as a link to Learn with Tiago for his awesome courses. And on the screen right now, you’ll see more certification content. Make sure you like the video, subscribe to the channel, and I’ll see you in the next one.

Copilot Studio + Planner Build an Agent That Actually Gets Tasks Done

Vlad Catrinescu — Mon, 24 Nov 2025 13:00:00 GMT

What if you could talk to your task list instead of clicking around in Planner?

In this video, I’m joined by Microsoft Senior Cloud Solutions Architect Damien Bird, who walks me through the task-planning agent he built in Microsoft Copilot Studio. This agent has no knowledge sources at all – it’s fully focused on doing work for him: creating, listing, and updating tasks in Microsoft Planner.

If you’re experimenting with AI agents, Copilot Studio, and Planner, this is a great example of how to move from “chatbot” to a real assistant that actually gets things done.

👉 Check out Damien’s channel @DamoBird365 for more Power Platform and agent content.
👍 Like, comment with the next agent you’d like to build, and subscribe so you don’t miss the rest of the PPCC interviews.

Want to learn from more pros? Learn from some of the top talent around the world at Microsoft Community Conference, and get $150 off with discount code VLAD150! Have questions? Check out my M365 Conference FAQ!

Watch my 100+ courses on Pluralsight

Video Summary

Agents are the big trend in 2025 – Instead of manually managing tasks in Planner, Damien built a Copilot Studio agent that lets him talk to his planner and create or update tasks conversationally.
No knowledge base needed – Unlike most demos, where you start by adding knowledge, this agent works without one because its purpose is purely action-driven: creating, listing, and updating tasks.
Smart orchestration with language models – The agent doesn’t rely on rigid conditions like Power Automate flows. It uses AI reasoning to match tasks even if you don’t type the exact name (e.g., “wash a car” vs. “wash the car”).
Voice commands and mobile integration – After publishing the agent to Microsoft 365 Copilot, Damien showed how you can use dictation on PC or mobile to create tasks by voice—like “Buy Vlad a beer tomorrow”—and it even understood “tomorrow” as a date.
Extensible beyond Planner – This approach works with any tool that has a connector, like Trello or even email drafting. If the connector supports due dates on creation, it’s even easier. The demo proved how fast and flexible building agents can be.

For more information, read the transcript blog below, or watch the video above!

Transcript

If there’s one thing that is popular in 2025, it’s agents. We all know that we should build them, but sometimes we wonder what kind of cool agent we can create to help make ourselves or our organization more productive. Luckily, today I’m joined by Damien Bird, who used to be an MVP and has now joined Microsoft to help lots of big customers make the most out of the Power Platform. Thank you so much for joining me today.

You’re welcome. Yeah, good to be here,” Damien replied.

So, today you told me you have a really cool agent you want to show me that helps you manage tasks. I’ll let you take it away—tell us what you’ve built, and then I’m going to ask you a bunch of questions as I learn about this as well.

Cool. Cheers, Vlad. First of all, apologies for the croaky voice, but I have my Copilot Studio agent already pre-built. It’s my task planner agent. I have my instructions at the bottom—I can even probably scroll. There you go, instructions at the bottom detailing some of the capabilities of this agent. But as you do with any agent, you can add knowledge. In this case, there is no knowledge because it’s ultimately an agent to create tasks for me on my Planner board.

I think this is the first time I’ve seen an agent with no knowledge. Every time I demo it, you create an agent and add knowledge—that’s always the first step.

Yeah. I built this purely to assist me with my Planner so I don’t have to do things manually using the Planner app. I’d much rather be able to talk to my Planner rather than do it manually. So, absolutely no need for knowledge in this case because I’m not asking it questions about SharePoint or Dataverse. It’s purely to create tasks, etc. I have these three actions already added. It’s just a case of adding them to tools and choosing your actions to add. These are the same actions you will see across the platform.

So you’re really using the connectors, not flows or anything like that?

Yeah, that’s right. In theory, you can pick and choose the ones you’re looking for—create a task, update a task, list buckets, which is quite an unusual one. I’ve added these to my agent. One of the key things to point out is the description, because this explains to the agent how to use that task, although that’s quite an easy description to understand. Then there are inputs, which can either be custom values—in this case, I’ve chosen to make the group ID and the plan ID custom values. That’s my group ID on my Planner and my particular plan ID because my plan is called ‘My Group Plan.’

So, in this specific agent, you hardcoded what plan everything goes into?

Yes. In theory, you don’t have to. If you’re happy to have that conversation with your agent and tell them that you want it to go into a particular plan, then if it can find it, it will do it. You’ll note that on the title, I’ve got customized settings—it’s set as dynamic. Again, description comes into play here. I’ve said the user might provide one or many tasks, identify each task, keep them short, and create a new task. That assists AI in understanding how to behave.

I have a question for you. I see you have here, ‘You do not need to ask them for a title.’ Usually, when it happens to m,e and I have something like that, it’s because it asks me for a title and then I have to specify, ‘No, you don’t need to ask me.’ Is this what happened here?

I’ve told you, yeah. Probably through iteration—when you’re building something, you go through the process of testing and then changing the prompt until you get the desired behaviour. So I’ve gone through a few iterations here and probably had that experience and said, ‘Right, you do not need to ask for the title.’ These inputs—some of them are dynamic. One of them is dynamic, the title. The other two are custom and fixed. I’ve done the same for the other actions as well. So I can create a task, list my tasks, and update a task. The reason I’ve got ‘update a task’ in there is that the particular action allows you to update the due date. When you’re creating these things, you want to look at the capability of the actions to see what they can do. A ‘create task’ action can’t set the due date, but an ‘update task’ can. So you have to combine these to create the experience you’re looking for.

That’s awesome.

In theory, I could have built a flow. I could have had a flow that defined some input parameters for a task that included a due date and combined those two actions in a flow, and then I’d have a purpose-built wrapper of these two actions to do that specific task. But I found that I could get away with it and talk to it. So that’s really all there is to it.

So you added the tools in your description. Did you add the name of the task or just let AI handle it?

If you’re looking for very specific outcomes all the time, then you can be very specific in your instructions. The instructions I’ve got on screen are based on the Copilot wizard experience at the beginning. You have a conversation, set up the agent, and it gives you some set instructions. I said at the beginning, ‘I want you to be a task planner agent,’ and it gave me instructions. Other than that, I haven’t spent hours manually crafting it.

No?

No. The way I look at this agent is I know the capabilities of the actions, I know it’s connected to my plan on Planner, and I’m just going to have a conversation with it and instruct it based on what I’m looking for. Every conversation can be different. Sometimes I might give it three tasks, sometimes one task. I might say, ‘Can you list all my tasks?’ or ‘Can you tell me all my due tasks in the next week?’ Those interactions are pretty much non-deterministic. I haven’t defined them anywhere. I can have that conversational experience, and it will do it. It will choose from those tools to get me that information.

For instance, I could say, ‘Can you list my current tasks?’, and what should happen is that the orchestration will look at my tools. It’s a live demo, so fingers crossed. There you go—it’s listing the tasks currently in my Planner. If I bring up my Planner, you can see I’ve got some tasks already in there. That’s what it’s surfacing.

I could say, ‘Can you update the wash a car task to the 1st of November?’ I’m curious because here you said ‘wash a car,’ but the name is ‘wash the car.’ Is it smart enough to know?

Wow. It did. Let’s see proof. It’s currently the 17th of October. I’ll refresh. Close your eyes for good luck. There we go—it did it. So smart enough to say, ‘Hey, it’s close enough.’ What’s happening is the orchestration activity map initially did a list task to get all the tasks back, then looked at the response data, and using the task ID, it updated that specific task. Even if I didn’t get it quite right, it’s not looking for that exact string like you’d have in traditional deterministic flows with condition statements. It’s using a language model to look for something similar in the output of the list task and say, ‘It must be that one—get the ID.

Copilot Studio shines more than just doing a Power Automate flow because Power Automate is deterministic—it must match, and even case sensitivity is something you hit in flows all the time. This uses a language model, and that’s the power of generative orchestration in Copilot Studio—the fact that you have a language model that can reason over the ask and look for the most likely outcome.

But now if I want to take it to the next level, you see now I’m typing. Can I talk to it?

Wow.

This is where I’ve published it to my Microsoft 365 Copilot. I discovered that yes, you can use dictation or even Windows dictation—the Windows + H key. I could say, ‘Can you create me a task to meet with Vlad on the 29th of October?’ Let’s send that. It didn’t get the date, unfortunately.

Is that because that would technically be two actions—it has to create it and then update it?

Yeah. So again, what I could do—let’s give it another try. Maybe ask it to update the task. ‘Can you please update the task to a due date of the 29th of October?’ We’ll send that across. Behind the scenes, it should look at those three tools, probably do a list of tasks, find the Vlad meeting, get the ID, and keep the context as well, because I didn’t mention which task. There we go—it’s updated. Cool.

That is pretty cool. Technically, you built it with Planner, but you could build this if your company uses Trello and there’s a connector for Trello. You could build the same thing for Trello.

Of course. It might be different depending on what the connector supports—maybe Trello lets you set the due date on creation, which would be easier. But you can do it with whatever task planning tool. 100%. You could even create draft emails or something on the go—anything you can bring a connector in for. You can now talk to it not only on your PC but on your mobile phone. I can fire up Microsoft 365 Copilot, turn on dictation, and just say, ‘Create me a task.’ Let’s try that as well. We won’t have the phone plugged in, but we’ll see in Planner if it works.

Right, I’ll get my phone out. Do you want to try it in your voice?

I can try. Right—‘Create a task to buy Vlad a beer tomorrow.’ Okay, we’re sending it. Come on. Then it’s in your task list. I’m getting an email. I got a beer. Okay. It says it’s being created. Let’s have a look. Yes, it ended up tomorrow—you got a date too.

Wow, that works. Thank you. How cool was that?

I’m going to love the reminder when you forget to delete it tomorrow.

No.

That is pretty cool, and it’s really cool to see how easy it is to build and how smart it can b,e where we said ‘tomorrow,w’ and it did it. That audio is built in—you didn’t have to do anything in Copilot Studio to activate the voice, right?

No. I could come back in here right now and ask, ‘When am I due to buy Vlad a beer?’ It picked up more than it needs. The agent is going to be like, ‘This guy’s weird—why is he talking like that?’ It didn’t find it. I wonder if any responsibility kicks in about buying alcohol—maybe that’s why. I’m surprised it didn’t find it. It happens—it’s a live demo. We’ll figure it out. But I could say, ‘What tasks have I got tomorrow?’ Let’s try that. ‘What tasks have I got due tomorrow?’ Come on—buy Vlad a beer. Yes.

Last day of the conference—I got three beers. Awesome. Delete that task.

No, you don’t have a delete action, do you?

No, I don’t actually. I think you’re stuck with that. But I could change the due date to next year.

Well, this was really awesome.

Cool.

You showed us how to do it in not even 20 minutes. We managed to test it quite extensively with two different accents on two devices. That was really awesome. For everybody who wants to learn more, you have a very popular YouTube channel. What is the name of it? Because I know it’s not your name.

It’s DamoBird365.

We’re going to have a link to it in the comments. If you want to learn more about agents, Power Automate, and the Power Platform—not only agents—make sure to check it out. Thank you so much for taking the time today to show us this. For everybody watching, if you want to see more cool stuff we recorded at the Power Platform Community Conference, check out the video appearing on your screen right now. And of course, make sure to like the video and subscribe to the channel. See you soon.

How to Use AI Builder Prompts in Copilot Studio (Demo)

Vlad Catrinescu — Tue, 30 Sep 2025 12:01:00 GMT

Want to see how to use AI Builder prompts inside Copilot Studio to transform data using one of the 11,000 models from Azure AI Foundry? This tutorial shows you exactly how to do that — and it’s straight from my full Pluralsight course 👇

🎓 Watch the full course on Pluralsight

In this video, you’ll learn how to:
✅ Build an AI prompt that pulls vacation data from Dataverse
✅ Choose from over 11,000 models, including GPT‑4.1, GPT‑5, and even Grok or Gemini
✅ Dynamically inject user-specific data into your agent’s responses
✅ Turn your Copilot Studio agent into a true assistant — not just a chatbot

Whether you’re building HR self-service agents, automating support, or exploring what’s possible with AI in Microsoft 365, this hands-on demo gives you the tools to get started right.

Video Summary

AI in Microsoft 365 is evolving — it’s not just about generating text anymore. It’s about creating intelligent agents that can take real action, like pulling live data to help users.
Copilot Studio makes it easy to build smart agents — in the demo, the speaker adds a topic called “vacation days” that responds based on user intent, not just trigger phrases.
Agents can connect to real databases — instead of giving generic info, the agent pulls personalized vacation day data from a Dataverse table, which could also be connected to systems like Workday or even Excel.
You can choose from thousands of AI models — over 11,000 models are available, including GPT versions and others like DeepSeek or Gemini, allowing you to tailor the agent’s capabilities to your needs.
The result is a truly helpful assistant — the agent dynamically retrieves vacation day info based on the user’s email and responds with accurate, personalized data, turning a simple chatbot into a powerful tool.

For more information, read the transcript blog below, or watch the video above!

Transcript

AI in Microsoft 365 isn’t just about Copilot writing text—it’s about agents that take action. In this demo for my Pluralsight course, I demonstrate how to use AI prompts in Copilot Studio to retrieve real data using your model. Let’s jump right in. I’m continuing to build our employee self-service agent, and for this next demo, I want to add a topic. I’m adding one from blank, and you’ll notice something different: the agent chooses based on intent. I don’t need a list of trigger phrases. I just explain what the topic does, and the agent will redirect the user based on that intent.

For this topic, I’m calling it “vacation days.” The description includes queries like “vacation days,” “how many vacation days do I have,” “check my vacation balance,” “PTO,” and so on. You can give it a list of phrases or just describe when to use the topic, like when the user wants to know how many vacation days they have left. It’s really up to you, and I recommend experimenting with your own scenarios because every model and situation is different.

Earlier in the course, I tested asking about vacations, and the agent gave general info from my knowledge sources—like “if you’ve worked between 0 and 3 years, you get this many days.” But that’s not good enough for my organization. I have a Dataverse table that logs how many vacation days each user has left. You might have this data in Workday or another HR system—or even in Excel (hopefully not, but it would still work). I want my agent to go into this Dataverse database, get the number of days I have left, and send me that info.

To do that, I’m using a new prompt from AI Builder. I name it “prompt vacation days.” When I go to the model, I see five default choices: GPT-4 mini, GPT-4.1, GPT-5, GPT-3, and GPT-5 Reasoning. But I can also add models from Azure AI Foundry, which has over 11,000 models to choose from. So if you want to use DeepSeek, Gemini, Grok, or others, you can. That’s amazing.

In my prompt, I tell it to search the database and add the table from Dataverse. There’s no search function, so I scroll to find my “employee holidays” table. I keep the user email and the “days left this year” columns. Then I add a text input called “user email” and use sample data—Vanessa, for example. I tell it to only return the number. When I test it, I get 35, which is correct. For this scenario, I probably don’t need a specialized model like Grok 3, which would add cost. GPT-4.1 works fine and also returns 35.

This shows how you can choose the right model for your business needs. Your agent can use 20 different models if needed, depending on their skills. I test it again with Vlad’s email, and it returns 50—perfect. I save the prompt and go back to the topic. Now I need to give it the dynamic user email. I create a new variable called “days left this year” and save it. That stores the info, but it doesn’t tell the user yet. So I add a message: “You have [days left this year] vacation days left.” I save and test it.

I type “vacation days,” and it hits the topic. It runs the prompt, and since I’m logged in as Vlad, it tells me I have 50 vacation days left. That’s awesome. Now the agent can look up info in a Dataverse table, in addition to all the knowledge it has from my intranet and benefits site. We’ve empowered it to connect to another system and help the user. And there’s even more we can do. That’s how AI prompts can turn a simple agent into a real assistant.

It’s been a wonderful journey—from building to publishing and monitoring agents. Check out the full course on Pluralsight by clicking the link on the screen or in the description.

Learning Basic Agent Creation in Copilot Studio

Vlad Catrinescu — Wed, 16 Apr 2025 12:00:00 GMT

Want to create your own Copilot Studio agent? This small clip from my Copilot Studio intro course on Pluralsight shows you the basics on how to create your first agent with Copilot Studio!

If you want to learn everything, check out the course here

Watch my 100+ courses on Pluralsight

Video Summary

Adding the SharePoint site URL to the agent’s knowledge base ensures clarity for future reference.
The agent creation process is quick, taking less than a minute, with markdown formatting making instructions easy to read.
Testing the agent with a question about the expense policy, which it successfully retrieves from the specified SharePoint site.
The agent displays the requested information along with the source link, ensuring transparency.
Initial setup is impressive, but there’s room for further optimization to enhance the agent’s performance and capabilities.

For more information, read the transcript blog below, or watch the video above!

Video Transcript

Now that we know the base architecture, let’s go ahead and create our first agent. Creating an agent is easy to do from the Copilot Studio web interface, but you do need to prepare a few things in advance to make this experience even faster. A few things you need to prepare before creating your agent are your agent name and the agent icon. I highly recommend that you work with your marketing and communications department to ensure they are visually appealing. Of course, you also need to provide an agent description so people know what they are using. So far, so easy. But after that, we have the agent instructions and the agent knowledge, which are a bit more advanced. So let’s dive deeper into them.

First, let’s talk about agent instructions. This is where you tell the agent what its role is, what it should do, how it should answer back, and whether it should be funny or serious. If we break it down, there are a few things you need to make sure that your instructions cover. First of all, the purpose: make sure that you clearly define the role and the area of expertise of the agent, including the primary functions that it serves. Next up, the guidelines: define how the agent should communicate. An agent might be concise, detailed, interactive, or suggestive. After that, tell the agent the main tasks that it is expected to perform. This can include generating suggestions, automating tasks, providing explanations, or guiding a user through a process. Finally, another important aspect is non-standard terms that the agent might not understand, such as any unique acronyms that only your organization uses. Instructions have a limit of 8,000 characters, and while that might seem like a lot, I have seen companies use it all. You can also use markdown to emphasize or organize your instructions. Markdown is a markup language that allows you to format plain text documents with special characters.

Let’s take a look at an example and start with the purpose. We can tell our agent: “You are an HR onboarding assistant designed to help new employees transition smoothly into their roles. Your goal is to provide essential onboarding information, guide employees through setup steps, answer common HR-related questions, and escalate complex issues when necessary. You should be friendly, professional, and efficient in all interactions.” Now let’s talk guidelines. The formatting, such as the hashtag, is markdown, but it is not mandatory. As you can see in the example on the screen, I tell my agent what tone to use, what to do, and what not to do. For example, “Do not provide confidential or personal employee information.” Next up, let’s talk about the skills, which are the main tasks that an agent is expected to perform. In this case, I tell the agent that it should retrieve and summarize HR policies, guide employees through IT setup like email, Teams, VPN, multi-factor authentication, and so on. I really tell my agent what tasks it will have to do when my users interact with it. Finally, I’m sharing some non-standard terms that my users or my documentation might use, such as “learning engagement unit,” or that “PS” means Pluralsight, or that “Excalibur” is the name of our intranet. Imagine if we never told our agent that Excalibur is the name of our intranet and somebody asked for a link to it; the agent would have no idea what they’re talking about.

Now let’s talk knowledge. What knowledge should this agent have? Should it have general AI knowledge, which is the public knowledge of the large language model on which it was trained? Should this agent access any public websites, any specific SharePoint sites, Dataverse tables, or can they even simply upload specific documents with knowledge for this agent? Now let’s go ahead and create our agent. Let me go create over here, click on “New Agent,” and the first thing I want to talk about is you see by default, the primary language here is English. If you want to edit it, you have to do it right now because once created, you cannot change the language of your agent. So let’s start giving it a name. The name of my agent is “Global Mantics HR Onboarding Agent.” You don’t need to add “agent” to the name. The best thing would be to work with your IT department and figure out what governance you want to put in place on naming, icons, and things like that. This way, all the agents that your company creates follow the same pattern. So you must work with IT as well as communications to pick the right icon, name, and things like that. This way, every agent you create makes sense for your users.

Now let me add a description here: “This is the Global Mantics HR agent that helps new hires with questions about Global Mantics.” And then my instructions. Now I will copy this; it’s what I showed you in the slides a bit earlier. If ever you want to try it as well, I’ll give you all the instructions that you can download as part of the exercise files. Let me put those in. There we go. You see, “You are an HR onboarding assistant for Global Mantics,” and really everything that we have covered in the slides. Something that I want you to notice: you see here I use markdown for headings, things like that, and here in plain text. However, after we create the agent, you will see that it understands it and formats it correctly. But just remember, here we use markdown, and I’ll show you the result after in just a few seconds.

After that, we need to add the knowledge. At creation, you don’t actually have all the different options. You can add a public website, you can add one or more SharePoint sites, and if we go in advance, you see we cannot add those yet. So you cannot add anything if you don’t want to, if the option you want isn’t there. But for me, I want to pick SharePoint, and I have my intranet here called Excalibur. Let me copy the URL of this SharePoint site here and click on “Add.” Great, it added the link, the name, and the description. I can, of course, optimize those, so if I go back to it or somebody else goes to edit it, they understand exactly what each knowledge source should be. Great, I clicked on “Add” here. You see, I have the Excalibur site; everything looks good. So let’s go ahead and click on “Create.” Now it should only take anywhere from 10 seconds to 1 minute, and wow, in this case, it only took less than a minute to create it. It’s still loading here, but there you go. What I wanted to show you in the instructions is that markdown is now formatted, so you see it looks pretty. That is a big advantage of using markdown in your instructions. When you consume it in the user interface, it will be a lot easier to consume.

Let’s do a quick test to see if it works. We haven’t optimized anything yet, right? We just gave it instructions and told it where to look. Let me type a question for my agent: “What is the expense policy for Global Mantics?” Let’s see if it goes to get it. We’ll give it a few seconds; we just created it, so it might not work right away. But there we go. The expense policy for Global Mantics is as follows: I get all the information, and then I also see where it came from. If I look at it here, you see it shows me the link to the file inside Excalibur, where it found that knowledge. For now, the only thing we have done is give knowledge and instructions to our agent. It’s already pretty amazing, but can we make it better?

TOP European Power Platform Conference 2024 Keynote Highlights | Copilot Studio, AI News & More

Vlad Catrinescu — Thu, 13 Jun 2024 11:42:00 GMT

Get all the top European Power Platform Conference (EPPC) 2024 Keynote highlights in Brussels from two Microsoft MVPs, Vlad Catrinescu and Gokan Ozcifci!

Watch my 80+ courses on Pluralsight

Video Summary

Here are the key points from the video:

Conference Overview: The European Power Platform Conference in Brussels had a great turnout with 2,000 attendees, doubling from previous years. The keynote highlighted the growth of the Power Platform community, now with 5.8 million monthly active users and over 600 user groups.
Product Highlights: The keynote covered various products, including Power Apps, Power Automate, and Copilot Studio. They emphasized the new mindset of Power Platform makers working alongside teams of copilot assistants.
Copilot Studio Demo: Despite a humorous interruption due to a Windows update, the demo showcased creating a chatbot on Power Pages that could handle both generic and specific queries, including order modifications.
AI and Automation: The introduction of AI flows and AI recording for desktop flows was highlighted. These features allow users to describe tasks to Copilot, which then builds the automation flow, making complex automation more accessible.
Future Vision: The speakers discussed the transformative potential of AI in app development, predicting that over 1 billion apps will be created by 2030 with AI assistance. They also emphasized the importance of managed environments for governance and administration.

For more information, read the transcript blog below, or watch the video above!

Video Transcript

Hello everyone! Goken and I are in cloudy Brussels today for the European Power Platform Conference. I know you don’t really agree with me about the “cloudy” statement, but it’s summertime for us. We just came out of the European Power Platform Conference keynote, and we are here today to give you the top highlights from the conference. So, Goken, what do you think about the conference so far? This is day one of the conference. We had a workshop together yesterday as well. What are your thoughts so far?

We are in Brussels, my hometown, so by default, it’s a 10 out of 10. We had an amazing keynote. We’ve seen a lot of stuff, especially with Power Apps. I was amazed by what we’ve seen, which we’re going to explain in a couple of minutes. There are a lot of delegates, phenomenal sponsors, and good weather. Don’t listen to Vlad; he says it’s cloudy, but look, we have the sun. You cannot expect more, right? The sun just showed up because you said it. You know how to talk to the Belgian sun.

Let’s start talking about the keynote. They started off by sharing some super interesting numbers with us. I know you took note of everything, so can you share them with us?

Yes, we have been told that we have 5.8 million community monthly active users and more than 600 user groups active today, which is phenomenal considering where it started. It started from nothing a couple of years ago, so to see that grow to 600 user groups worldwide and over 6 million people answering questions on the forum is just amazing. I feel like we’ve also experienced that in attendance at the Power Platform Conference as well. We’ve both attended almost all the Power Platform conferences so far, and in the keynote, they shared the evolution of attendees at this conference. The first one was in Berlin three years ago with 500 people. We then went to Dublin, which had 1,250 attendees, so we doubled. Now, we are at 2,000 in Brussels within three years, which is fantastic. I think it’s 2,000 because they sold out. We heard that quite a while ago; the workshop sold out over a month ago. It’s crazy that we could have had even more attendees if there was space.

Now, Goken, let’s start talking about products. What’s the first product they talked about on stage?

We talked about a lot of stuff, from Copilot Studio to automation to Power Apps to managed environments. They started the keynote with the mindset that’s changing within the Power Platform community as well as the platform. They told us that we had Power Platform makers, and this year we are Power Platform makers plus copilots. Now, they are making the platform to be Power Platform makers plus a team of copilots. So, we don’t only have one copilot; we’re going to have our little copilot teams that help us out. It makes me think, “Minions, assemble! Come help me make an app.” Probably that’s going to be the case. I love it because we also see that in SharePoint. You can build your own copilots, and the way they go, I feel, is the same, which has not always been the case. I love it, to be honest.

Do you love it? I love it. If you love it, I love it too. Then we went to Copilot Studio, which went great. They had to show us three main pillars: the agent, knowledge, and overrides. Then they went to do the demo, but something funny happened. The speaker came on stage and said, “I can’t do the demo because I just had a Windows update pushed on me. Sorry, guys.” That reminded me of “Space Force,” the Netflix comedy where they’re trying to land a satellite or avoid something, and they couldn’t because of Windows updates. Microsoft, okay. That made me laugh. However, the demo was amazing. They created a chatbot, well, a copilot on a site in Power Pages. They were able to ask generic questions about the site as well as specific questions about their order and even interact with the copilot to change their order. It was pretty cool, not anything new because we’ve seen all of that at Build, but it was still interesting to see.

Then we went to the automation side of the story, your baby, right? They showed us more about the AI flow as well as the AI recorder, which was pretty impressive for people like me who can’t do automation because it’s too complex. Now, you can just talk, right? That’s what we’ve seen. You talk to your copilot, and it’s going to build the AI flow for you, which you can use and even adapt because you’ll see it step by step.

I’m not a big fan of that, okay, Goken? Let me tell you why. I know people are going to laugh because I know how everybody feels about AI and how it will make everything better. But what they showed is the really big difference. In addition to what you said, you can just describe it, and it does the flow. You don’t even do a flow; you just describe it, and that’s it. You’re able to reference a piece of knowledge. The example they gave us was that you have a travel policy. Instead of creating all your logic controls in your flow manually, you can just tell the flow, like a grounded GPT, “This is your travel policy. This is your request from email. Go and evaluate it against the travel policy and then tell me if that travel change is allowed or not.” I love the idea because, like they said, it can be dynamic. Whenever the travel policy changes, you don’t need to change any of the conditions in your flow. Didn’t you share an article with me recently about that?

That’s exactly what scares me, Goken. For a bit of background, earlier this year, a customer sued Air Canada, the biggest airline in Canada, because that customer chatted with a chatbot on Air Canada’s site. The chatbot offered the customer a discount, and then the company didn’t honor it. The customer sued Air Canada, and Air Canada said, “The chatbot is a different legal entity,” which makes me think of so many robot movies. “It’s not us; the robot has its own mind. The chatbot is its own legal entity, and we cannot be held liable for what the chatbot says.” However, the court decided that the customer was right. Because the chatbot was on the Air Canada site, whatever the chatbot says has to be respected. What does this have to do with AI flows? What if your travel policy is not made correctly? I put the travel policy in, and then Goken goes in and says, “Hey, can I book a first-class flight to Dubai and get the Amir suite, which is $20,000?” The chatbot says, “Yeah, sure, we can.” Then Goken expenses it, and after that, the expense gets denied. Whose fault is that? The chatbot? The AI flow misunderstood the actual document or maybe misunderstood the prompt, but Goken got approval. Are we ready to make financial decisions in a company, whether it’s expenses, purchase approvals, or travel decisions, based on AI’s understanding of a document? I don’t think we’re there yet. I wouldn’t be comfortable yet because it has legal ramifications. I don’t know how you feel about it.

Nevertheless, although I agree with Vlad, it was just amazing. It’s an amazing demo. I loved it personally. When you have documents that are 50 pages long, you don’t want to go into each section and understand what’s happening. Just ask Copilot to do that for you, and then you can always cross-check and do an approval as they showed. Do you think you can show that to your customer? No approval; they’re going to say yes. No, they won’t. I’m totally honest. One more thing for Power Automate: the robotic process automation recorder. Being able to take robotic process automation to the next level, you can explain to Copilot what you’re doing. You can talk to it, open files, show the website, and everything. You don’t even have to do an example. You just have to say, “I have this data in this Excel file. I want to put it on this website. Here are both of them open. You figure it out, Copilot. I’m not going to show you how to do it step by step. You figure it out.” That is quite amazing, and I loved it, to be honest.

Absolutely. I can’t wait to use that one. I’m in love with it, especially since I don’t like doing RPA flows now. That will save me a lot of time. Absolutely. After that, we switched over to Power Apps, which is your baby. We got a demo with a lot of examples of how you could integrate it with Copilot. The first one was the creation of the canvas app, where the presenter just wrote some apps. Instead of creating the app with one table, we went to data modelling and had multiple Dataverse tables with sample data. The fantastic part was that there were already references and lookups into those data tables. Within Copilot, you could also add more tables with sample data and make references with lookups. Fantastic stuff that I personally never do, like data modeling. I always build my app first and then go for the data modeling, which is a bad thing, to be honest. Now, within Power Apps, you can do it from the start.

One of the things we said before is that you can’t use Copilot to create your data because it only creates one table. Now, it’s able to create relationships and truly create everything, so that’s crazy. Please, if someone at Microsoft listens to us, I know databases are cool, but please give us the chance to also choose other data sources such as… okay. Then we had a demo of co-authoring. When the presenter created this canvas app, we could clearly see someone else working on the app and adapting the layout, as well as stuff that’s already in our tenant, like “Explain me this formula” or “Create me a formula with Copilot.” Pretty good, actually even a tremendous demo for those who have never seen Copilot in action with Power Apps, especially data modeling sites. I’m really looking forward to using it. Awesome. And then we ended on managed environments. Managed environments, managed environments. I feel like Microsoft is adding everything they have that has to do with administration and governance. Everything is going to managed environments. I feel like if you’re a big company, you don’t have a choice now but to use managed environments for third-party tools because that’s the only way you can get governance features. And how many? Ten. Ten new features. Ten new features that Microsoft announced for managed environments. I’m not sure how many of them are new at this conference versus Build, but there are at least ten upcoming features in the coming months for managed environments.

One thing they said before the closing is how AI Copilot will change the way we make our apps. Two things: first, with Copilot, things that were impossible before are now becoming possible because we get more and more intelligent and can do more stuff. Secondly, almost 40% of the apps are now AI-infused or using Copilot, which is true. I have a few apps, and the first thing I did was, “Hey, let me see if I can do it better with Copilot or if I can add more stuff with Copilot,” which is a phenomenal statement and which I 100% agree with. I don’t know about you, but I think it will change the way we do things. They even shared that people are going to build over one billion apps. We should say it like that: one billion apps by 2030. Okay, there we go. We’ll see who’s right after when we put the overlay, but over one billion apps, even if it’s by 2030, is still a huge number. So having AI help us create them and being able to infuse AI into them will change the way we work.

But I think that’s it for the highlights. Those were the top highlights. Thank you, everybody, so much for watching. I truly hope you enjoyed it. Make sure you come to the next European Power Platform Conference or the Microsoft Power Platform Conference if you’re more in North America and want to go to a local conference. Microsoft even talked about the Microsoft Power Platform Conference, which I found quite funny. But if you do, make sure you use code VLAD100 to save $100. And if not for the next European Power Platform Conference, please sign up early. They’ve been selling out almost every year, so please sign up early to make sure that you have your spot. We don’t know yet when it’s going to be or where it’s going to be. My guess? I like Hawaii, so hopefully they go to Hawaii. I’m more realistic; I say Prague. I would love to go back to Prague. But we’ll see. We’ll see who’s right. But thank you very much, everyone, for watching. Again, make sure you like and subscribe to the channel for more awesome videos like this.