Who really benefits from the AI agent boom?

Hint: It's not you.

Every week, there’s a new promise: AI agents that will replace your team, run your business while you sleep, and make you a millionaire by Tuesday.

But if AI agents were really this powerful, you’d see results everywhere. Not just in social media posts.

Unfortunately, the reality is simple: AI agents aren’t transforming businesses. They’re creating the illusion of progress. And that illusion is profitable. As long as we all keep falling for it.

Because while we’re all clicking, sharing, and hoping for shortcuts, the real winners are the companies and the creators selling the agent story.

Here’s what’s really going on…

Billions have been poured into large language models that still don’t make money for the companies behind them (for example, ChatGPT’s growth has been stratospheric, but 95% of its customers use the free version). And it will take many billions more to pursue Artificial General Intelligence - which is what these companies are really after.

AI companies need to fund that. And agents offer a potential solution.

They give AI companies a way to say, “Look, this is the next frontier.”

They give influencers a way to say, “I’m ahead of the curve: click here.”

And they feed our FOMO with the fantasy of agentic AI: the story that overnight success is possible. And it can be automated. But it remains a fantasy.

Hey Reader. If you like this post please like ❤️ this post - it helps tell SubStack that it's worth sharing far and wide. 

Four uncomfortable truths about AI agents

1. They’re not easy to make.
No matter what the influencers say, building an AI agent that actually works is complex. Yes, there are tools that make it fun to experiment, but safely and effectively deploying one (not sure that’s even possible right now) takes far more than drag-and-drop.

2. They hallucinate.
You can’t give a large language model arms, legs, eyes, and ears, then set it loose while it still makes things up. Obviously. These systems get things wrong. Sometimes spectacularly wrong. No intelligent business leader is going to risk their reputation, clients, or data on that. Right? Right???

3. Most “agents” aren’t agents at all.
When an influencer tells you about the agent they have built that is transforming their life and business, mostly what they mean is that they have built an automation: a rule-based workflow that they rebrand as an agent (it’s called agent-washing: old tech, new label, same limitations).

4. They’re risky and unpredictable.
Even when they work, they can misfire, loop endlessly, or act on bad information.
They need guardrails and human oversight, that often undoes any promise of time saving.

AI agents aren’t the revolution you’ve been promised. They’re a fascinating experiment - one worth understanding - but they’re not ready to run businesses, replace teams, and they’re not making anyone millions (besides maybe the AI companies that are hyping them like mad).

What’s happening right now is a story being sold. And like all good stories, it works because we want it to be true.

That said, the smart move isn’t to reject AI agents altogether. It’s to understand them for what they are, experiment safely, and keep humans firmly in the loop. That’s where real progress happens.

Want to know how to intelligently incorporate AI into your business? Register for my workshop on 6 November: The Small Business AI Adoption Strategy.