We’ve all seen what generative AI can do—but AI agents are here to take productivity to the next level. In this clip, part of my Using AI Agents for Productivity course on Pluralsight, you’ll learn how AI agents can act on your behalf, learn from experience, and integrate with your tools to help you get more done.
🎓 From the course: Using AI Agents for Productivity
Watch my 100+ courses on Pluralsight
Video Summary
- AI chatbots are great—but limited: Tools like ChatGPT, Microsoft Copilot, and others have boosted productivity, but they don’t know your company data, can’t remember past interactions, and often miss the mark on your personal or organizational standards.
- Prompt engineering is exhausting: Writing long, detailed prompts every time you need something done is time-consuming and inefficient. It often feels like more work than just doing the task yourself.
- Enter AI agents—your proactive digital teammates: Unlike traditional AI, agents don’t just wait for commands. They understand your goals, take initiative, and get smarter over time by learning from your preferences and actions.
- Three key features make agents powerful:
- Roleplaying: Each agent has a defined role, making it laser-focused on specific tasks.
- Memory systems: They remember your preferences and past interactions to tailor their help.
- Tool integration: Agents can take real action by connecting with your organizational tools.
- Real-world examples show the magic: Imagine a social media agent that creates and schedules posts in your brand voice, or a project manager agent that updates tasks and meeting notes automatically. Even field techs can have support agents that order parts and access past ticket data instantly.
For more information, read the transcript blog below, or watch the video above!
Transcript
Generative AI tools have increased our productivity and made many aspects of our life easier. But as we are entering the era of AI agents, you may be wondering: can agents make our lives even easier and enable us to be more productive? My name is Vlad Catrinescu, and in this course, we will learn about AI agents in productivity scenarios.
First of all, why do we need agents? I think you’ll agree with me that for the past few years, the focus has been all on AI chatbots—from the popular ChatGPT to Microsoft Copilot, Google Gemini, Anthropic Claude, and even social networks creating their own tools and models with Grok AI and Meta AI. And don’t get me wrong, those tools are awesome, but they have some limitations.
First, they’re all grounded in the internet, meaning that their knowledge all comes from the same data, which is publicly available on the internet. They do not have access to your company data, and they do not know you or remember who you are from one interaction to another. And even those tools that do, they often have trouble alternating between different needs, such as asking different types of format for your different clients or projects. And because of the combination of both those things, the answers you get don’t always follow your organizational standards or your personal objectives. It’s always a bit of a hit or miss.
And okay, I have to admit, there was one way to get answers always adapted to you. And I’m sure many of you have seen prompt engineers on different social media sites show you those huge prompts, like the one here on the screen, where basically the only thing I’m asking this chatbot is to give me some social media posts. But because I have to tell it everything, it ends up being more like a novel. And honestly, it might be shorter to just write those social media posts myself than to write huge prompts like this every single time.
Again, even if you wrote those long magic prompts, you still always had to take that first step and write the prompt. Nothing was proactive. But for a second, imagine an assistant that’s not just passively waiting for your commands but actively working alongside you—an AI tool that understands your goals, makes decisions on your behalf, and learns from its experiences to become more effective over time.
While traditional Gen AI focuses on producing content based on prompts, AI agents are designed to take action. They can perform tasks, interact with different systems, and even collaborate with other agents or humans to achieve complex objectives.
There are three main features that agents have that enable them to become more effective. First of all, roleplaying. When using agents, each agent is assigned a specific role or persona that guides its behavior and interactions. This isn’t about pretending or simulation—it’s about defining a clear scope and context for that agent’s operation. This makes agents laser-focused on what you want each agent to do.
Next up, we have memory systems. Memory systems enable AI agents to retain and recall information from past interactions. And this isn’t just about short-term memory—it’s building a knowledge base that the agent can draw upon over time as it gets to know you. Whether it’s the way you want something written or your meeting scheduling preferences, such as you don’t like to have meetings before 10:00 a.m. even if your workday starts at 8.
Finally, and probably the most important, is tool assignment. AI agents are designed to take action, and they are integrated into all your organizational tools that you have access to.
So what are some types of agents that we can create? First, since we had a social media example before, let’s imagine a social media agent. This agent would have been trained on all your previous social media posts, campaigns, and metrics. It also knows your brand guidelines and information related to your goals, such as your upcoming event. You simply need to ask it what to do, and it will create those social media posts in your voice, your guidelines. It could even create images based on your templates that you might have in Adobe Creative Cloud or in Canva. And because it’s integrated with your social scheduling tool, once you approve them, it will automatically schedule them at a time that would maximize interaction based on your specific metrics and past performance.
Another example could be a project manager agent. This one—you don’t even have to do anything. It just fixes your meetings, and it does everything from creating meeting notes to updating tasks in your project management tool based on your conversations. You agree to delay a task while talking to your colleague? Done. It’s updated. Something is done and you haven’t updated it yet in the project plan? It will mark it as done for you. It’s the virtual assistant you always needed.
You could also have a technician support agent for your on-the-road techs that is trained not only on your whole documentation but also on all the previous tickets and resolutions ever recorded for that product. It also knows your inventory and it’s connected with your shipping platform. So if something is broken, it will directly order that missing piece for you.
