Bolted On vs. Built In

Two firms, same subscriptions

Two professional firms of similar size bought the same AI tools this year. Both have the assistant, the meeting summariser, the document tool. Both spend roughly the same. One of them now carries thirty percent more client work with the same headcount. The other operates exactly as it did last year, with more logins.

The difference was not the technology. Both firms rent the same models, from the same handful of providers, at the same price. The difference is where the AI sits relative to the work.

What bolted on looks like

Bolted-on AI lives beside the work. Someone opens a separate tab, pastes in context, gets an answer, and carries it back by hand. It knows nothing about your clients, your matters, or the decision your firm made on something similar three years ago. It is a very good general-purpose assistant that happens to be near your business.

This is why adoption stalls. Two enthusiasts use it constantly. Everyone else tries it, finds the effort of explaining context each time roughly equal to doing the task, and quietly stops. Nothing is wrong with the tool. It simply has no way to become part of how work moves.

What built in looks like

Built-in AI sits inside the workflow, on the firm's own record. The question is not asked in a separate tab; it is answered where the work already happens, from the firm's clients, matters, deals, and history. The reasoning is visible, so an answer can be checked rather than believed.

That changes the economics. A bolted-on tool saves an individual some minutes. A built-in system changes what the firm can carry, because the work moves without waiting for a person to shepherd it between systems.

The test

There is a simple way to tell which kind you have. Ask what happens if you cancel the subscription tomorrow.

If the answer is that a few people go back to doing something manually, it was bolted on. If the answer is that your operations stop, it was built in, and the second thing is the only one that was ever going to change the firm.

Why most firms end up with the first

Because buying is easy and building in is not. A subscription needs a card. Built-in AI needs someone to map how work actually moves, decide which workflow is worth starting with, connect systems that were never designed to talk to each other, and get people to trust the result. That work is unglamorous and it is the entire job.

The firms pulling ahead did not find better AI. They did the part everyone else skipped.