The Cognitive Wall is real. But it's not where you think it is.

May 25, 2026 · 2 min read

Most enterprises I speak to believe they’ve hit a ceiling with AI. Pilots stall. “Strategic insights” turn out to be confident hallucinations. The data science team is buried. And somewhere, a competitor is supposedly pulling ahead.

Here’s the uncomfortable diagnosis: the wall isn’t the AI model.

An AI Enterprise seat doesn’t give a company an AI strategy any more than buying Excel gave it a finance function.

Claude Opus 4.7, GPT-5, Gemini 3, these systems are extraordinary out of the box.

The wall is everything wrapped around them. Context engineering. Evaluation discipline. The willingness to redesign a workflow instead of bolting a chatbot onto it. The judgement to know when not to use AI at all.

The companies quietly winning aren’t the ones with the biggest AI budgets. They’re the ones who treated the model as a commodity and invested in the surrounding craft — taste, governance, and a clear definition of “good” before a single pilot launched.

The ones who hit the wall? They bought seats, ran pilots, measured nothing meaningful, and concluded the technology wasn’t ready.

The technology was ready. The organisation wasn’t.

A few uncomfortable truths I’ve been sitting with

Most “AI failure” is governance failure in a trench coat. Internal data science teams are sometimes the bottleneck, not the solution.

“Falling behind competitors” is the most over-sold anxiety in enterprise software right now. A lot of those competitors are also faking it.

Being a thoughtful late mover is genuinely defensible.

So I want to throw it open

Where have you actually seen the Cognitive Wall in your organisation — was it the model, the people, the process, or the politics?

And the harder question: what would you do differently if you started your AI strategy from scratch tomorrow?

Curious to hear from the people in the trenches.