Most businesses think they've cracked AI. They've built a brochure website.

"We're not looking to deploy new AI tools".

I'm hearing this more regularly from executives, and every time it tells me two things.

The first is understandable, but it's reading the problem wrong. The second is what needs solving. This is brochure thinking.

When the internet arrived, most businesses treated it as a new channel. A website was a digital brochure. Email was a faster letter. Useful, but it missed what the internet actually was.

The internet was an application layer. The brochure was just the surface. Underneath it, entirely new capabilities were emerging that couldn't have existed before: self service portals that eliminated call centres, supply chain coordination across continents in real time, marketplaces that disintermediated entire industries, analytics that turned every customer interaction into a feedback loop. Many of these took decades to materialise as mobile, cloud, and integration platforms matured, and most businesses struggled to see beyond the surface to imagine what else was possible.

The difference with AI is that the technology is already mature enough to build on, the infrastructure doesn't need a decade to catch up, and there are more leaders, frameworks, and precedent to draw from. The gap between seeing the possibility and acting on it has never been smaller.

"We don't need another tool" is brochure thinking. It treats AI as a product category that sits alongside the CRM and the analytics platform. That framing caps everything before anyone's even started.

The underlying issue isn't too many tools. The tools don't talk to each other, so the human becomes the integration layer, manually stitching the full picture from fragments across a dozen platforms. Consolidating from ten tools to five just creates fewer, deeper silos.

Disconnected systems mean no one has context. And context is everything when it comes to AI. The entire value of an AI system comes down to the right information, at the right time, in the right place, to make the right decision. The outcome is only ever as good as the context it's working from. Siloed tools mean siloed AI: individually intelligent, contextually blind.

The application layer is what happens when AI connects every system into a single intelligence layer. CRM data enriches marketing decisions, marketing performance feeds back into operations, operations insights inform finance forecasting. Each system makes every other system more valuable.

The Head of Ops starts their day with priorities ranked across every system. The Marketing Director gets campaign performance cross referenced against segments, purchase data, and margin. The CFO interrogates a live model of the business instead of reconciling spreadsheets.

Same people, same tools, completely different capability. And there's a progression to get there.

One person, one task, one prompt. Nothing connected to business systems. Output gets copy pasted back into whatever platform they were working in. Nothing compounds.

The CRM predicts churn, analytics surfaces anomalies, marketing optimises send times. Each tool is contextual with its own data but blind to everything else. This is where most businesses sit, and where the "we don't need another tool" instinct comes from. Rational, disciplined, and capped.

AI sits between the tools, not inside any one of them. An insight in one platform triggers an action in another, automatically. Same stack, no new tools, but the human stops being the integration layer. The gap between insight and action compresses from days to minutes, but the underlying processes and structures remain the same.

AI changes what the work is, not just how fast it gets done. Instead of reacting to data, the system anticipates. Instead of humans deciding and AI executing, AI recommends and humans steer. Roles shift from execution to orchestration. The org chart changes shape because the work has transformed. This requires genuine organisational change: culture, structure, roles, decision making.

Nobody jumps from Level 1 to Level 4, but understanding where the business sits changes the question worth asking.

At Level 1, the question is whether the AI capabilities already built into existing platforms are being used. Most businesses are paying for features they haven't turned on.

At Level 2, the question is where humans are being the glue between platforms that should be talking to each other. That's the path to Level 3, and it doesn't require new tools.

At Level 3, the question is which decisions are still being made manually that the system could handle, and where the business is organising work around constraints that no longer exist.

They're right, they don't.

Tools were never the point. The point is context: how much of the picture does anyone see when they make a decision, and how fast does insight become action.

Most businesses are at Level 2, thinking about Level 1 problems. The ones that pull ahead will see the full map and start moving.

Written by Mike ✌