AI is everywhere. Your sales team is using ChatGPT. Marketing is experimenting with prompts. Someone in RevOps has built a spreadsheet with a model bolted on the side. On the surface it looks like progress.
Underneath it is a more uncomfortable number.
That is a commercial failure rather than a technical one. Most businesses have adopted AI perfectly well. What they have yet to do is operationalise it.
AI is being used:
- Individually, where it should be organisational
- Tactically, where it should be strategic
- In tools, where it should be in systems
The result is scattered usage that produces output and never builds into advantage. Everyone is busy and the business is no better off.
Why “everyone has ChatGPT” changed the game
For a brief moment, access to AI felt like an advantage. That moment has passed. Models are commoditising, capabilities are converging, and access is abundant.
The advantage was never using AI. It is where AI sits in how your business makes commercial decisions.
In most businesses, value is created and captured in a small number of places: go-to-market decisions, pricing and packaging, sales execution, revenue operations, and customer expansion and retention. Where AI reaches those, it shows up in results. Where it stops at the individual tab, it stays a productivity story.
This is why so many AI initiatives stall at pilot. They optimise tasks, and the value lives in systems.
What happens when AI lives in personal tools
Knowledge fragments
Insight sits inside individual sessions and goes when the tab closes. The reasoning behind a good call this week is unavailable to the person making a similar call next week.
Answers vary
Two people ask the same question and get two different answers, both plausible, neither grounded in what the company actually knows.
Nothing carries forward
Decisions stay separate from each other. There is no loop between what was decided, what happened, and what gets decided next, which is the gap that separates the AI investments returning something from the ones that do not.
The commercial impact stays invisible
Activity is easy to see. Its connection to pricing, go-to-market execution and revenue is not, so it never reaches a board conversation.
Tools, and an operational layer
A tool assists an individual. An operational layer changes how decisions are made, how work flows, and how value builds across the organisation. Getting to the second takes four things:
- Shared context, so everyone is answering from the same picture
- Defined workflows, so the answer lands somewhere it can be acted on
- Governance and visibility, so the reasoning survives the decision
- Feedback between action and outcome, so the next decision is better informed
Without those, AI stays at the edge of the business.
This is a commercial design problem
Most AI projects that failed had good models, capable tools and enthusiastic teams. What they lacked was a place in commercial decision-making, in execution rhythms, and in accountability structures. AI was added on rather than built in.
Which is the same reason most go-to-market strategies fail. They live in decks rather than in operations.
The companies seeing a return are doing something different. They design AI into the commercial workflows that matter, treat it as a system rather than an experiment, measure value creation and value capture explicitly, and let insight build over time. They are running smarter commercial operations.
From usage to advantage
The shift is deliberate. Personal usage becomes organisational systems, prompts become processes other people can run, and the experimenting turns into advantage.
Left to itself, AI stays a productivity story and somebody else takes the opportunity. The companies that win this phase will be the ones that design it into how value is created and captured.
That is the real work.