There’s an addiction sweeping through the business world right now, and nobody’s calling it what it is.
Every week, a new tool promises to “automate your workflow” or “replace your team with AI agents.” The pitch is always the same. Set it up, let it run, watch the magic happen. And thousands of business owners are buying it, literally, without stopping to ask a fairly obvious question: does any of this actually make you more productive?
I’ve been testing OpenClaw extensively over the past few weeks. For anyone unfamiliar, it’s the open-source AI agent that went viral in January, racking up 145,000 GitHub stars in under two months. It runs on your desktop, connects to your tools, and promises to handle tasks autonomously. It’s genuinely impressive technology.
Here’s the problem. To run it with models that are equivalent to (or better than) an average team member, you’re looking at API costs in the thousands per month. Not tens. Not hundreds. Thousands. One power user reported spending $3,600 in a single month after his assistant consumed 180 million tokens managing his calendar, emails, and task lists. Another burned through $200 overnight on what they described as “simple scheduled tasks.”
For that money, you could hire a brilliant executive assistant. Or two. Train them to use AI as a tool (not a replacement), and you’d have someone who understands context, reads the room, builds relationships, and exercises genuine judgement. The kind of things that no autonomous agent, however sophisticated, can reliably do.
The Automation Frenzy is Real, and the Data Backs It Up
This isn’t just my observation. Research published in Harvard Business Review tracked 200 employees at a US technology company over eight months and found something uncomfortable. Workers with AI tools didn’t work less. They worked at a faster pace, took on a broader scope of tasks, and extended their hours, often without being asked. The tools didn’t free them. The tools made them busier.
Economists have a name for this. The Jevons Paradox: when technology makes a resource more efficient to use, total consumption of that resource tends to rise, not fall. It happened with email. It happened with spreadsheets. And it’s happening now with AI.
Meanwhile, McKinsey’s own data shows that nearly 90% of companies have invested in AI, but fewer than 40% report measurable gains. Over 60% see no significant bottom-line impact. The gap between investment and results isn’t a technology problem. It’s a thinking problem.
Superhumans, Not Replacements
I’m not anti-AI. I run a data-driven marketing agency and we use AI tools daily. But there’s a meaningful difference between using AI to supercharge a capable human and trying to replace that human with an autonomous agent that costs more, understands less, and needs constant supervision.
The businesses getting real value from AI aren’t the ones automating everything. They’re the ones giving their best people better tools. An account manager who uses AI to prep for client calls in half the time. A marketing coordinator who uses it to analyse campaign data that would have taken days. A finance lead who uses it to spot patterns in cash flow that they’d have missed manually.
That’s augmentation. That’s where the value is.
The automation-for-automation’s-sake crowd is building elaborate Rube Goldberg machines while their competitors are simply training their people to be twice as effective. And the irony? The people building those machines are spending so much time maintaining them that they’ve created more work, not less.
The Real Question
Before you spend another penny on autonomous AI agents, ask yourself this: would this money be better spent making the people I already have significantly better at their jobs?
Nine times out of ten, the answer is yes.
The future of work isn’t humans versus machines. It’s humans with machines, guided by people who know when to automate and, more importantly, when not to.
