From my perspective, the suitability report is the single most time-consuming document in a mortgage or IFA brokerage. Every recommendation needs the client objectives captured, the affordability narrative, the lender rationale, the Consumer Duty evidence, the risk warnings — and it has to be in your firm’s tone, not generic. Most advisors I speak to spend 60 to 90 minutes per report. For a busy brokerage that’s 15 hours a week on one document category.
I believe this is one of the highest-value Ai applications in regulated financial services, specially because the gains are so concrete. Drafting time drops from 90 minutes to 15, without loss of quality, with a cleaner audit chain than manual drafting produces. But only if the Ai tool is built to handle client data correctly — which most aren’t.
Why most Ai tools fail this workflow
The core problem is that a fact-find contains the most sensitive client data in your file — NINOs, dates of birth, income figures, debt balances, health information. To draft a suitability report, the Ai needs to see at least some of this. Pasting it into ChatGPT is a non-starter under FCA rules. Uploading it to Copilot, Gemini or Claude has the same problem — you’ve sent personal financial data to a third-party model that may train on it.
The compliant workflow needs three properties. One, the client data stays inside your firm’s tenant. Two, the Ai that drafts sees redacted data — placeholder tokens instead of real NINOs, for example. Three, every interaction is logged to an audit chain the FCA can examine. Without all three, you’re either slow (manual drafting) or non-compliant (general Ai).
The 15-minute workflow, end to end
Specially in governed platforms built for FCA-regulated practices, the workflow looks like this. Five minutes to upload or import the completed fact-find into the case vault. The Ai reads it, flags inconsistencies, pre-populates the client-objectives section of the report. Apart from this, it cross-references the fact-find against your firm’s suitability template — which risk warnings are required, which disclosures apply.
Five minutes to generate the first draft. The advisor enters the lender recommendation and the rationale; the Ai drafts the affordability narrative, the stress-test commentary, the Consumer Duty plain-language section, the risk warnings. The advisor reviews section by section on screen.
Five minutes to edit and sign. The advisor edits for tone or clinical accuracy, runs the plain-language check, signs the draft. The audit chain captures advisor ID, timestamp, source data fields, draft versions, sign-off. Report is sent to client the same afternoon. Previously that same report would have been drafted overnight and sent Thursday.
The PII redaction that makes it work
From my perspective, the non-negotiable piece is the redaction layer. A governed Ai platform should strip NINOs, DOBs, income figures, bank details, debt values — replacing them with placeholder tokens — before any content leaves your tenant for the Ai model. The model sees “[NINO]” instead of the real number. The model drafts around the placeholder. When the output comes back, the placeholder is restored on your side.
This is how Other Me handles it. You can read more about the architecture on the PII handling post or, for the specific FCA workflow, on the Mortgage and IFA solution page.
SYSC 9 audit: what the chain looks like
The audit chain from a governed workflow is stronger than the audit chain from manual drafting. Every fact-find field the Ai read is logged. Every draft version is versioned. Every sign-off is timestamped with advisor ID. If the FCA asks “reconstruct the advice you gave Mr & Mrs Hargreaves,” you click the case, export the chain, done.
We’ve written separately on SYSC 9 record-keeping requirements for Ai-assisted advice. Apart from this, for Consumer Duty specifically, every client communication auto-attaches the metadata — plain-language score, fair-value evidence, comprehension check — that your Consumer Duty Champion needs at month-end.
What to do this week
If you’re a principal at a mortgage or IFA brokerage, try one suitability report in a governed Ai platform this week. Pick a real case. Compare the time and the audit chain to your current workflow. If it’s material, roll it out to the advisors. If it isn’t, you’ve spent an hour learning something.
You can start a free 7-day trial, no credit card. The trial is the full product, so you can draft a real report in it, test the PII redaction, examine the audit chain. If your compliance lead wants to review the architecture before the trial starts, the Security page publishes the technical documentation openly — no NDA, no sales call.