Whilst the impact of over $200 billion being wiped from enterprise software and later insurance market valuations might suggest AI is taking over the world, it isn’t. In the right hands, AI is a powerful tool that can materially complement value creation in software and insurance. But let’s dig into where we actually are.

$30 to $40 billion invested in artificial intelligence over the past two years. MIT reports 95% of organisations saw zero measurable return. Not reduced return. Zero.

PwC surveyed 4,450 CEOs for Davos: 56% report neither revenue gains nor cost savings. Only 12% say AI delivered both. Forrester now predicts a quarter of planned 2026 AI spend will be deferred to 2027. Kyndryl found 61% of CEOs face more pressure to prove ROI than a year ago — and 65% aren’t aligned with their CFO on long-term AI value.

IBM’s global managing partner put it plainly: organisations “sprayed and prayed rather than systematically asking how the technology would make the company better.”

That word — systematically — is doing all the work in that sentence.

Ninety-five percent of organisations did the same thing. Took an unchanged operating model. Bolted AI onto the surface. Called it transformation. Copilots on unchanged workflows. Chatbots on unchanged processes. Agents on unchanged architectures. The pattern is consistent. AI bolted onto unchanged operating models — features on the surface, nothing changed underneath. Fragmented pilots without governance — marketing launched a chatbot, finance tried forecasting, IT piloted automation, none connected.

The 5% producing real returns did something fundamentally different. They redesigned the system so AI was intrinsic — woven into how the product thinks, not just what it does. Learning-capable systems that compound with every interaction. Workflow redesign, not decoration. Compound intelligence where every data point strengthens every other.

I introduced a name for this to market advisor thinking in global P&C insurance: systematic AI. Not a feature set — an architectural substrate. Intelligence lives in the database. Agents are stateless workers. The knowledge graph is the brain. The 5% aren’t using different tools. They’re using the same tools differently — connected through architecture, not sprayed across surfaces.

The slides walk through the evidence, the diagnosis, the pattern, and what the 5% are actually doing. I’ve been operating in AI for over a decade — not the current hype cycle, the long arc. The methodology is validated across enterprise insurance, autonomous fintech, and intelligent relationship platforms. Not prototypes. Working systems. The architecture transfers because it was designed to.

Bolt-on is a bandaid on a broken leg. Systematic AI is the surgery.

If your organisation is part of the 95% — or if you’re building for the 5% — the conversation starts with architecture.

I’m easy to find.

#AI #EnterpriseTransformation #ProductLeadership #CPO #InsurTech

First Comment

The IBM quote deserves its own moment. “Sprayed and prayed” — that’s the global managing partner of one of the world’s largest consulting firms describing how most of his clients approached the biggest technology shift in a generation.

The pattern in the slides is deliberate. Bolted-on AI. Fragmented pilots. Outputs measured instead of outcomes. Every enterprise I’ve seen stall has at least two of those three. Most have all of them.

The 5% redesigned the architecture first. That’s not a feature decision. It’s a structural one.

I’ve spent a decade building exactly this — enterprise insurance, autonomous trading, relationship intelligence. Same methodology. Different domains. Each one compounding. 151 patent claims across 15 inventions. The architecture transfers because it was designed to.

The conversation starts with architecture. It always does.