Working with AI and big documents: Getting results so reliable that you can run your business on them

A lot of business still runs on big documents. A customer contract checked against your own terms. An employment agreement checked against the award. A tender or security questionnaire answered from everything you’ve said before. Each one needs an exact answer rather than a fairly good representation, near-enough synthesis or fast summary. Businesses have invested in AI tooling but are often finding the results aren’t quite what they hoped for - and having to revert to manual analysis anyway.

More of this work now sits with the teams who own it, using tools like Claude without an engineering team behind them. The results often look right and turn out not to be, and someone redoes the job by hand.

I’m running a “BigDocs” session for legal and commercial, HR, sales and marketing, support and operations teams, and the CTOs who want those teams to self-serve. I’ll walk through the patterns from the BigDocs series and the benchmark behind them, including where my own tools fell short, and we’ll look at real examples from the room.

If your team has a document job that comes back from AI looking right but turns out not to be, bring it.

We’ll diagnose what’s happening, how to fix it, and give you the practical skills to sort it out.