Artificial intelligence does not change what good commercial architecture looks like. It changes how quickly the consequences of bad commercial architecture appear.
This is the part of the conversation that most businesses are not having — and the part that will matter most over the next three years.
The tools are being adopted. That is happening at a pace that is genuinely difficult to keep up with, and the outputs they produce are compelling enough that the adoption is accelerating rather than slowing. Sales teams are using AI to write outreach. Marketing functions are using it to produce content at a volume that was previously impossible. Finance teams are using it to model scenarios that used to require a week of analyst time.
What is almost entirely absent is the governance layer — the structural framework that defines what AI is allowed to influence in the commercial system, how those outputs are reviewed, and what architecture they are being applied to.
The risk is not that AI produces wrong outputs. The risk is that AI produces convincing outputs from a commercial system that was never designed to produce the right ones.
I have worked with sales organisations that implemented AI outreach tools and saw activity metrics increase dramatically. More sequences running. More touchpoints generated. More pipeline entered into the CRM. The conversion rate on every stage declined. The total revenue generated did not move.
What happened was structurally predictable. The AI accelerated the output of a commercial motion that was misaligned — the positioning was not resonating with the buyers being targeted, the sequences were reaching the wrong people at the wrong stage, the value proposition was not differentiated enough to produce a response. The AI made it possible to do all of that more frequently and at lower cost. The result was more misaligned outreach, executed more efficiently.
AI governance in a commercial context is not a technology question. It is a commercial architecture question. The relevant questions are not about which tools to adopt or how to prompt them effectively. They are about whether the commercial architecture being accelerated is strong enough to produce the outcomes being attempted — and whether the organisation has the structural discipline to review what AI is producing before it compounds into the market.
The businesses that navigate this well are not the ones running the most sophisticated AI stacks. They are the ones that examined their commercial architecture first — positioning, channel alignment, pricing discipline, signal quality — and then applied AI to accelerate a system that was already producing the right direction.
The businesses that navigate this poorly are the ones that added AI to a commercial system they had never examined, and discovered that the thing AI is exceptionally good at is making existing problems larger, faster.