An established law firm had adopted AI early. Over a few years it built up a mix of tools: some made in-house, some by an outside developer, some bought off the shelf. Each one solved a problem when it arrived, but nobody had stepped back to look at the whole picture.

The firm’s principal didn’t need convincing that AI works. He needed an honest view of which tools were earning their keep.

He also wanted to know what to build next, and how the firm’s older core systems would hold any of it back. R6AI ran a half-day working session with the principal and the people who run the firm’s systems. No pitch deck. We mapped a matter from first enquiry to final bill, then reviewed every AI tool against how the firm actually works.

The answer was focus. Most of the value sat in three tools, so those get deepened instead of the firm adding a tenth. The rest are kept as they are, wound down deliberately, or parked.

The bigger finding sat underneath the tools. The firm’s own records, built up over many years, were its most valuable AI asset. Their quality was the first thing to fix, because every future build would inherit it.

The firm left with a findings report, a full map of its AI tools and systems, and a staged build plan that starts small and proves itself early. It also left with clear criteria for choosing its next core platform, so that decision gets made on fit, not on a sales demo.

The session didn’t add another tool. It showed the firm where the value already sat and built a plan to protect it.

Every build follows the same rules: client data stays protected, a person reviews anything before it leaves the firm, and AI never does the maths.