The File That Takes Three Days
The move in short
Custom Build — A pipeline triggered when a new client completes their onboarding questionnaire and fact-find submission that pulls the form responses, any uploaded statements or documents, and the client's stated goals and risk profile, runs an LLM synthesis to produce a structured pre-meeting briefing pack — financial snapshot, key concerns flagged verbatim, suggested agenda, and a list of clarifying questions — and delivers it to the assigned adviser's inbox 24 hours before the first meeting.
The Company
Meridian Wealth Advisers is a fee-only IFA firm based in Edinburgh, about 30 people. They work with high-net-worth individuals across Scotland — business owners, retirees, people who've had a liquidity event and suddenly have more to manage than they expected. No commissions, no product sales. Clients pay for advice, so the quality of that advice is the whole business.
The Pain
Claire is the operations lead. When a new client finishes their onboarding, she gets a form submission, maybe a few uploaded statements, and a notes field where the client has described their situation in their own words — sometimes a sentence, sometimes three paragraphs. Her job is to make sure the adviser walking into that first meeting knows what's in front of them.
So she reads through everything, pulls out the relevant numbers, flags anything that looks complicated, and writes up a short briefing. Then she does it again for the next client. When they're onboarding two or three people in a week, she's spending most of her time on this. The advisers rely on it. It's genuinely useful. But it doesn't scale, and Claire knows it.
The Move
They'd build a small automated pipeline triggered the moment a client submits their onboarding questionnaire. The pipeline — something like Make or Zapier — picks up the form responses, pulls in any uploaded documents, and passes everything to an LLM with a consistent prompt that asks it to produce a structured briefing: a financial snapshot from whatever the client has shared, any concerns flagged in the client's own words, a suggested agenda for the meeting, and a list of clarifying questions the adviser should probably ask. That briefing lands in the assigned adviser's inbox 24 hours before the first meeting.
The tools for this exist now. Make and Zapier both handle document retrieval and can call the OpenAI API directly. The prompt takes some care to write well, and they'd want Claire to review the output format against a few real examples before they switch it on. But there's no custom software here. A few days of setup, and Claire's no longer the one doing this by hand.
The blind spot
The onboarding form already exists. The documents already get uploaded. Everything needed to generate this briefing is already arriving in one place. Because the inputs are all there, it doesn't feel like an automation problem — it feels like reading and summarising, which feels like a human job. So nobody thinks to hand it off.
The pattern
The same approach fits a few other situations:
- A consulting firm where a project manager has to brief new team members on an existing client before they join a call
- A recruitment agency preparing a candidate summary for a hiring manager before an interview, drawing from a CV, application form, and job spec
- A law firm where a partner needs a quick read on a returning client's file before a call they haven't had time to prepare for