Now apply that concept to artificial intelligence. Instead of responding to questions like ChatGPT or Siri, an agentic AI system works independently. You give it a goal and it researches, decides and takes action without you having to micromanage every step. Bill Gates put it this way, “Agents are not only going to change how everyone interacts with computers. They’re also going to upend the software industry, bringing about the biggest revolution in computing since we went from typing commands to tapping on icons.”

We’re at the beginning of this shift. Right now, AI agents can manage emails, schedule meetings and even run financial models. In the future, they could negotiate deals, handle customer service or even plan your retirement. It’s a fundamental change from today’s AI, which mostly waits for you to tell it what to do.

If you’ve used ChatGPT, you’ve experienced a powerful AI — but one that responds to prompts rather than acting on its own. It pulls from a vast pool of training data and gives you a well-reasoned answer, but it doesn’t do anything unless you tell it to.

Agentic AI, on the other hand, operates with autonomy. It sets its own tasks, makes decisions and learns from interactions. Instead of just answering, it might schedule appointments, draft reports and follow up with people — all without your constant oversight.

Think of it like hiring a good assistant instead of using a search engine. You don’t have to look up a dozen flight options, weigh the pros and cons and book the best one yourself. You just say, “Find me the best flight under $500 that lands before noon,” and the agent handles the rest.

Right now, most AI agent platforms are built for businesses, but there are ways to try one yourself.

1. AI agent marketplaces — While there’s no dominant marketplace yet, a handful of platforms let you pay for AI agents that can automate tasks, like customer support or marketing. These are evolving rapidly, but most are still aimed at companies rather than individuals.
2. Build your own — Microsoft offers AutoGen, an open-source framework for creating custom AI agents. It’s free, powerful, and integrates with Microsoft’s paid cloud services like Azure. While it lacks a user-friendly interface, it’s one of the best tools for building sophisticated agents that can tackle complex workflows.
3. No-code automation — If writing code sounds intimidating, simpler tools like Zapier let you automate tasks without programming. While not true AI agents (they follow set rules rather than making decisions), they can handle repetitive work like sending emails, updating spreadsheets and managing notifications. And in some cases, that’s preferable — AI agents are still unpredictable, and handing over too much decision-making too soon can backfire.

Take Klarna, the Swedish fintech giant. It wanted to cut its customer service team in half by using AI agents. Instead of reducing costs, the transition was a disaster — customers were left frustrated, and Klarna had to backtrack. AI agents, while promising, aren’t foolproof. They can make mistakes, and when they do, fixing those mistakes can take longer than doing things the old-fashioned way.

Over time, expect AI agents to become easier to use, with dedicated marketplaces full of reviews and recommendations — just like today’s App Store or Chrome Web Store. While most current tools are designed for businesses, consumer-friendly AI agents are on the way, built for everyday needs like personal finance, health care and travel.

For now, “build your own” solutions are popular among tech-savvy users, but as adoption grows, that approach will likely fade. In the same way most of us don’t build our own apps, we won’t need to create our own AI agents — just pick one that works best from a growing list of available options.

The key takeaway? Agentic AI is coming, and it’s going to change how we interact with technology. Whether that’s for the better or worse depends on how well these systems are designed — and how wisely we choose to use them. If you’re interested in learning more about AI, Stanford University has just released a resource full of free classes at https://craft.stanford.edu/resources on a broad range of topics including “How do I fact-check AI search results?” and “​​How can AI be used to help conserve natural ecosystems?”

Leslie Meredith has been writing about technology for more than a decade. Have a question? Email Leslie at asklesliemeredith@gmail.com.