Built around it, not bolted on

There’s a tell that separates companies that will compound their AI advantage from those that will spend three years and a lot of money standing still. It isn’t budget, and it isn’t talent. It’s where AI sits in the decision to do the work at all.

Bolt-on companies start from the existing process and ask: where can we insert AI to make this faster? The process is sacred; AI is a garnish. You get a copilot in the support queue, a summariser in the CRM, a chatbot on the website. Each one is real, each one saves a little time, and none of them change the shape of the business.

Built-around-it companies ask a different question: if we were starting this function today, knowing what a frontier model can do, what would we never build?

The question reframes the org chart

When you ask the second question honestly, the answer is rarely “the same team, slightly augmented.” It’s “half the steps don’t need to exist.” A workflow that took four handoffs becomes one agent and one reviewer. The reviewer is now the most important role, and the three handoffs are gone.

That is a threatening sentence in most companies, which is exactly why incumbents resist it and AI-native challengers don’t — they never hired the three handoffs in the first place.

The competitor you’re worried about isn’t doing your process faster. They’re running a process you’d be embarrassed to propose in your own all-hands.

What this looks like in practice

• Start from the outcome, not the task. “Resolve the customer’s issue” survives the rethink. “Route the ticket to the right queue” usually doesn’t.
• Make the human role explicit and senior. Built-around-it doesn’t mean unmanned. It means the human does judgement, not transcription.
• Measure against an AI-native baseline, not last year’s numbers. Beating your own prior quarter is comfortable and irrelevant if a challenger is operating at a different cost structure entirely.

None of this requires a moonshot. It requires picking one workflow, asking the harder question, and being willing to act on the answer. That’s where every engagement we run ends — with one thing actually rebuilt, not a deck about rebuilding everything.