You may have heard the word “agent” or “agentic” when talking about AI, especially with the new ChatGPT agentic model that was released in June 2025. But what does “agentic” actually mean?
Merriam-Webster defines agentic this way: someone or something capable of achieving outcomes independently (“functioning like an agent”) or possessing such ability, means, or power. When you are talking about an “agentic AI”, that means it’s not just chatting with you, but is empowered as an agent to take action in the real world. An agent can do things like navigate websites, filter results, prompt you to log in securely when needed, conduct analysis, and even deliver editable slideshows and spreadsheets that summarize its findings.
When demonstrating the new ChatGPT agent model, Sam Altman and his crew at OpenAI showed how it could do a complex task like making travel bookings. Another great example was ordering cupcakes for an office party by giving the AI agent the dietary requirements and budget for it to research and order the cupcakes.
We think a key thing to note, however, is this: agentic AI is coming quickly, but it’s not quite there yet. To go back to that cupcakes example, it took the AI an hour to order the cupcakes, and it didn’t get everything right. Ultimately, the person who created that experiment said, “you know, it would have taken me five minutes to do this myself.” Our recommendation is to use agents for tasks that they’re well-suited to, like rewriting content. One simple example for your business could be having an AI agent check your email and summarize what’s in your inbox or compose a draft email on your behalf to answer basic questions that have come in while you are offline.
Want to learn more? Check out our podcast: Episode #1 Practical AI Tips
(art by Becka Rahn)
