I’ve spent the last few posts talking about how AI is reshaping partnerships, marketplace economics, and GSI relationships. Every one of those arguments rests on the same assumption, so I want to state it directly. AI is the best tool I’ve used in thirty years of building software and running go-to-market teams. It is not an autonomous replacement for the people doing the work, and the data backing up the hype cycle is a lot weaker than the headlines suggest.
Gartner expects more than 40 percent of agentic AI projects to be scrapped by 2027. MIT’s research on generative AI pilots found that 95 percent failed to deliver measurable business impact. Those aren’t numbers from skeptics trying to slow the industry down. They’re coming from the same analyst firms and research institutions that have been documenting the upside all along. The gap between demo and production is real, and it’s bigger than most vendor pitches admit.
Why the gap exists
The math is straightforward once you see it. If an AI agent is 85 percent reliable at each step of a task, that sounds impressive in isolation. But string ten of those steps together, the way a real workflow requires, and the end to end success rate drops to roughly 20 percent. Models are genuinely strong at tasks that take a person a few minutes. They get noticeably shakier as a task stretches into hours and requires holding context across many steps without a human checking the work.
That’s not a knock on the technology. It’s a description of what the technology actually is: a tool that makes a skilled person dramatically faster, not a system you can walk away from and trust to run itself. Nearly nine in ten organizations running AI agents have already had a confirmed or suspected security incident tied to one, and more than half of deployed agents operate with no logging or oversight at all. The failures aren’t happening because AI is bad. They’re happening because people are deploying it as if it doesn’t need the same judgment, review, and accountability every other business system has always required.
What this means for how I use it, and how I sell it
In every conversation I’ve had about channel, marketplace, and GSIs, the throughline is the same. AI removes the mechanical drag from work so people can spend their time on judgment, relationships, and the calls that actually require a human to be accountable. A proposal draft, a first pass at a statement of work, a summary of a messy data set: hand it to the model. Deciding whether a customer is really ready for a deal, whether a partner relationship is healthy, whether a transformation program is actually on track: that stays with a person, because the tool isn’t built to hold that judgment across a long, ambiguous, high stakes sequence, and the failure data says as much.
The bet I’d make
The companies that win this decade won’t be the ones that tried to remove humans from the loop fastest. They’ll be the ones that used AI to make their best people more productive while keeping a human accountable for every decision that actually matters. Treat it like what it is: the sharpest tool you’ve ever had in your hand. Sharp tools still need someone skilled holding them.