88% of organisations use AI. Only 7% have scaled it.

And the gap is widening every quarter.

That gap — between AI activity and AI capability — is the single most important number in enterprise AI right now. McKinsey put it on paper in December 2025. We’ve been living inside it for six months, running operating-model reviews with three Australian enterprises that all hit the same wall.
Different industries. Different sizes. Different starting points. Same wall.
Here’s what the wall looks like up close.

The tooling zoo
Walk into any of these orgs today and you’ll find this:
• Marketing has ChatGPT Enterprise.
• Sales has Gong + Claude.
• Engineering has Copilot + Cursor.
• Ops has three agent platforms in pilot.
• Legal is reviewing six tools simultaneously.
• Nobody owns the portfolio.
Each team has a defensible reason for the choice they made. None of them are wrong. But the org has accidentally fragmented its AI spend, governance, and capability uplift across functions — and nobody is asking the question the board needs answered:
What’s in. What’s out. What’s funded. What’s killed.
This is the tooling zoo. It’s not a strategy problem. It’s an operating model problem. And it’s a fair chunk of why 88% can claim AI usage while only 7% can show compounding capability.

The reframe: AI Enablement is a function, not a programme
The orgs that close the gap stop treating AI as a programme and start treating it as a function — the same way they’d think about Finance, or HR, or Risk. There’s a portfolio to manage. A platform to run. Governance to enforce. Enablement to deliver. And measurement to defend in front of a CFO who does not care about prompt engineering.
Five pillars. One cadence. One team owning the question.
Without that, you get pilot theatre — endless demos, no compounding. With it, you get capability — pilots that ladder into platforms, platforms that ladder into governance, governance that ladders into measurable, board-defensible advantage.
This is the move we keep watching the leaders make. It is, almost without exception, the move the laggards have not made.

Seven traps that quietly cede ground
The tooling zoo is one of seven patterns we keep seeing. The others, in shorthand:
1. The tooling zoo — fragmented portfolio, no owner.
2. Pilot theatre — endless POCs, none productionised, board patience expiring.
3. Governance gridlock — every tool blocked for “review”; real decisions never made.
4. Shadow AI — the org is using AI. You just don’t know where, or what data is in it.
5. The champion-only trap — one excited team, no enterprise leverage.
6. The adoption vanity metric — “75% of employees have used it once” tells you nothing.
7. The measurement vacuum — no honest ROI number, so AI loses every reprioritisation conversation.
Each one is fixable. Most aren’t fixed because nobody’s been given the operating-model authority to fix them.

What the next 90 days actually looks like
The work isn’t a two-year transformation programme. It’s a sequence of operating-model moves you can stand up inside a quarter:
• Name the owner. One throat to choke. Not a committee.
• Build the portfolio register. What’s funded, what’s piloted, what’s killed — and a quarterly cadence the exec team actually attends.
• Pick one enablement spine. The same fluency programme for everyone, regardless of which tool.
• Stand up a governance frame that says yes faster than it says review.
• Define a measurement model that survives contact with the CFO.
None of these are tech moves. They are operating-model moves. Which is exactly why technology functions struggle to do them alone — and why the orgs that do them quickly are the ones with executive sponsorship and a named function doing the work.

Where you actually sit
Most orgs we talk to sit at one of five stages: Ad Hoc → Coordinated → Operationalised → Embedded → AI-Native.
You can’t skip stages. You can pick the right next stage and run at it hard. The mistake we see most often is trying to leap from Ad Hoc to Embedded by buying a platform — and ending up back at Ad Hoc, with a more expensive licence bill and a CFO who’s now suspicious.
Pick the right next move for your stage. Make it. Then make the next one.
That’s how the 7% become the 7%.

Walk through it live
We’re running the full session — all seven traps, the 5-stage maturity model, the 90-day execution playbook, the five pillars — twice in the next ten days.
• Mon 8 Jun · 4:00 PM BST — EMEA
• Tue 9 Jun · 2:00 PM ET / 11:00 AM PT — Americas
45 minutes plus 15 of Q&A. No moderator, no panel. Bring the hard questions: board pushback, governance deadlock, the CFO who won’t accept your ROI numbers, the team that’s gone shadow on you, the platform decision you can’t unwind.
Save your seat →
Can’t make either? Register anyway. Every registrant gets the recording and the 5-stage maturity scorecard within 24 hours of the live session.
— Audrey & Peter