TL;DR: Soft facts such as client doubts, family influence, health concerns, and emotional reactions are increasingly recognised as important compliance evidence, not optional colour. Capturing soft facts for your FCA file works best with a hybrid workflow: AI transcription preserves verbatim client language and timestamps, the adviser reviews and filters during a human review stage, and a final sign-off locks the file. Verbatim notes can protect the file better than polished paraphrase because a reviewer can trace statements back to what the client said.

For example, a client might say they are "not sure" about accessing their pension at 55. That hesitation is a soft fact, and if it is not on the file, the suitability rationale has a gap. This is the documentation problem sitting underneath the admin problem: 43.3% of UK advisers report that paperwork and admin reduce the time they devote to advice itself, and the detail most likely to be lost in rushed notes is exactly the detail an FCA file review will ask for.

The FCA's own supervision leadership has emphasised that capturing soft facts helps tell the client's story and provide evidence that advice is suitable. This article sets out what counts as a soft fact, why it matters for evidencing, and how to structure AI meeting notes so soft facts are captured, flagged, and aligned with the advice you give.

Defining Soft Facts for Your FCA Audit Trail

Defining Hard vs. Soft Client Facts

Hard facts are typically the quantifiable elements of a fact-find: age, income, existing assets, pension values, ATR score. Soft facts are the qualitative and contextual elements that explain the "why" behind a client's decisions, including personal and family health issues that may be relevant to the advice being provided.

The distinction matters because soft facts add context and help tell the story, and can demonstrate how your advice is right for the client. One case cited in that speech: a client's hereditary heart condition was a main motivation for the advice but was not recorded anywhere on the file, despite the firm believing the advice to be suitable. That is a defensibility gap, not a note-taking style choice.

Evidencing Client Needs in AI Notes

AI meeting notes capture soft facts as evidence rather than as loose transcription, provided the tool produces structured output rather than a flat text dump. Evie generates structured notes from client meetings, and its contextual understanding handles financial terminology, UK dialects, tone, and reactions, so the minute detail of a meeting lands in organised output. This is what capturing soft facts in AI meeting notes looks like in practice: the client's own words, categorised against the structure of your file.

Evidencing Client Needs for the FCA

The regulatory hook is direct. COBS 9 requires firms to take reasonable steps to ensure a personal recommendation is suitable based on the client's knowledge and experience, financial situation, and investment objectives, and to provide a suitability report explaining why the recommendation is suitable. Suitability documentation should evidence the reasoning, not just the conclusion. Under Consumer Duty (FG22/5), firms must deliver good outcomes, and the FCA has emphasised that evidencing Consumer Duty compliance requires data, not just processes. A file recording only hard data may struggle to demonstrate consumer understanding. The soft facts are where understanding lives.

Categorising Soft Facts for FCA Compliance

Not every soft fact carries the same weight. Five categories consistently matter in file reviews.

Capturing Emotional Context for FCA Files

Fear of loss, confidence in markets, and comfort with volatility all inform whether a recommended risk level is genuinely suitable rather than merely tolerated. Per FCA thematic review findings, including TR15/12, failure to establish whether the client can financially bear losses is a recurring suitability deficiency, and emotional context is part of how you can evidence capacity for loss alongside the numbers.

Capturing Client Doubts for FCA Files

Expressed concerns about a recommendation, requests for alternatives, and preference conflicts are among the most valuable soft facts you can hold, because they show the advice was tested against the client's own reservations. A doubt raised and addressed in the meeting notes evidence is far stronger than a file that reads as if the client agreed with everything.

Evidence of Family-Led Financial Choices

FG21/1, the FCA's finalised guidance on the fair treatment of vulnerable customers, reinforces the need to understand the circumstances driving decisions, and family dynamics are frequently part of that picture.

Recording Client Life Goals

The same TR15/12 thematic review identified inconsistencies between portfolios and the client's investment objectives and investment horizon as a key source of unsuitable recommendations. Life goals are the antidote to the generic objective.

Evidence for FCA-Compliant Advice

Health and longevity concerns cut across all of the above: life expectancy assumptions, hereditary conditions, mobility, and care planning directly affect product suitability and withdrawal strategies. If a health concern motivated the advice, it belongs on the file in the client's own words.

The Hidden Cost of Omitting Soft Facts

A Clearer Audit Trail

The cost of omission is usually invisible until a file review, a complaint, or a PI claim surfaces it. Advisers relying on memory and manual notes capture soft facts inconsistently, and the inconsistency compounds across a busy review calendar. Firms that have adopted AI meeting notes report a clearer audit trail on compliance files, with richer evidence captured per meeting reducing the time spent explaining files during external reviews.

Closing the Gap to Accurate Notes

The gap between what was said in the meeting and what reaches the file is where risk accumulates. The two gains aren't separate. Richer capture is what makes the time saving sustainable: output only requires editing, not reconstruction, when the underlying note is complete enough to trust. Atlas also reviews the transcript after the meeting unprompted, flagging missed referrals, data mismatches between the back office and what the client said, and records that need updating before the file is locked. That proactive pass means the gap is reviewed at the source, not discovered at the next annual review or during an external file check.

The Risk of Prioritising Data over Context

A file heavy on data and light on context invites exactly the deficiencies the FCA cites most often. When 71.9% of firms spend 1 to 7 hours producing a single suitability report, the temptation is to record the numbers and move on. The numbers alone cannot prove suitability. Context can.

Using AI to Document Nuanced Client Insights

Why Verbatim Notes Protect Your File

Some firms already use recording to capture client meetings, and verbatim transcripts can preserve the client's actual words as evidence. SYSC 9 requires firms to retain relevant client records appropriately, and introducing a third-party AI processor does not reduce that obligation. Advisers should also review consent and data handling obligations before deploying an AI notetaker. Timestamping can add audit trail value by demonstrating when something was said and what context existed at that moment, and a desk-based review is only as strong as the digital file submitted.

Tracking Emotional Cues in Meeting Notes

Here is the honest trade-off. AI transcription captures what was said verbally, but it has limitations in interpreting non-verbal cues, body language, or unspoken hesitation. Some errors can come from limitations in capturing nuance such as sarcasm, tone, or inflection, and AI notetakers can hallucinate to fill in gaps. A hallucination in this context means transcript text that misrepresents what was actually said. The adviser adds the observational layer during review, and that layer is non-negotiable.

Meeting Notes That Satisfy FCA Audits

After the notes are drafted, automated checking closes the loop. Evie and Colin are capabilities within Atlas, AdvisoryAI's AI chat and intelligence layer for UK advice firms. Colin runs 42 automated checks on a suitability report, producing a colour-coded pass/fail report with a percentage compliance score and specific remediation guidance. The checks cover AML documentation, client profiling completeness, risk assessment adequacy, recommendation suitability, and report quality, and Colin is system-agnostic: it checks any suitability report, meeting note, or fact-find, regardless of which system produced it. A defensible advice file requires every statement to be traceable back to its source, not just a log of the prompt and output. Atlas's Adaptive Thinking makes each step of its reasoning visible and persistent, so every query against your meeting data shows what was searched, what was loaded, and the reasoning behind the answer, and older queries remain auditable. AdvisoryAI's approach to governance emphasises outputs grounded in the firm's own data with source citations, gated by mandatory human review.

Structuring AI Notes for FCA Files

A robust workflow for capturing soft facts has three steps:

1. AI transcription: Record the meeting via Teams, Zoom, or Google Meet. Meeting notes and email summaries generate directly in Atlas, in your firm's own template, editable inline and downloadable as PDF or Word, or sendable via Outlook.

2. Human review and filtering: Review transcripts for accuracy, being alert for misattribution, misinterpretations, ambiguities, and fabrications, documenting corrections and context. This is where your professional judgment filters what belongs on the file.

3. Back office population and sign-off: Structured notes populate fact-find fields in Intelliflo, Iress Xplan, Plannr, or Curo, and the adviser or compliance team signs off before the file is locked.

Turning Client Anecdotes into Robust FCA Evidence

Client Uncertainty About Pension Access

A client who says "I'm not sure I want to touch the pension yet" has given you a soft fact that may shape drawdown strategy, tax planning, and review frequency. Captured verbatim with a timestamp, it can evidence that the recommendation accounted for the client's stated ambivalence rather than overriding it.

Capturing Spousal Risk Disagreements

When one spouse is comfortable with volatility and the other is not, that disagreement may be relevant for any joint recommendation. Recording who said what, and how the disagreement was resolved or accommodated, can protect the file if the arrangement is later questioned by the less comfortable party.

Documenting Client Health Concerns

Health details need careful handling. A hereditary condition that motivates protection advice belongs on the file, an unrelated medical anecdote does not. The adviser decides during review, and firms should consider deleting raw audio promptly once the file note is finalised, in line with data protection principles.

Documenting Client Hesitation and Body Language

Body language from audio-only recordings is one category AI cannot capture, and pretending otherwise undermines the whole file. Note it yourself during the human review stage: "client paused for an extended period and looked to their partner before agreeing" is a sentence only you can write, and it may be the most defensible sentence in the note.

How to Document Soft Facts for Robust FCA Files

Ensure Note Accuracy Post-Client Meeting

Treat every AI output as a first draft and review carefully before signing off. Because notes generate in Atlas using your firm's existing template, the output you are reviewing already matches your established document structure rather than a standardised vendor format. The time economics make this realistic: Brooks Macdonald reports meeting note time dropping from 2.5 hours to 30 minutes per meeting in an annual review workflow, and Timothy James and Partners cut post-meeting documentation time by 50%, with support teams accessing notes significantly faster.

LOA pack review time reduction based on Finsource Partners outcomes using Emma.

Flagging Soft Facts for FCA Compliance

During review, flag the soft facts that carry compliance weight: emotional context, doubts, family influence, life goals, and health concerns. Evie's structured output organises these details, and the Intelliflo integration supports fact-find updates directly into the client record.

Align Client Insights with Advice

Every flagged soft fact should connect to a recommendation or an explicit reason it did not change the recommendation. That linkage is what turns meeting notes evidence into suitability evidence, and it is what a file reviewer reconstructs when testing whether the advice was suitable for this client, not a generic one.

Guidance for Capturing Client Nuance

Use this checklist when evaluating or configuring an AI note-taking tool for soft fact capture:

• Data hosting: Confirm UK or EEA data residency, since advisers should ensure AI providers comply with UK GDPR data protection rules, including storage limitation and any international data transfers.

• Encryption: Consider encryption in transit and at rest, and confirmation that conversations are kept out of third-party training data.

• Consent process: Announce to all participants that an AI notetaker will be used, confirm consent, and document it on the record. Have a clear process for when clients refuse recording, with a manual note fallback.

• Human-in-the-loop verification: Mandatory adviser review before any note reaches the client file.

• Source citations and timestamps: Every statement traceable to the transcript, with the transcript traceable to the recording.

• Retention policy: Configurable retention, with raw audio deleted once the file note is finalised.

• Back office mapping: Soft facts map into structured fields in your back office, not a standalone document store. Transparency obligations run through all of this: meeting participants should be informed about AI note-taking, including what data is collected, where it is stored, and how it is processed.

Request a demo to see how Evie captures soft facts in your meeting workflow, using your firm's own templates and a real client scenario. There is no commitment: trials run for 14 days with no credit card required, contracts are monthly rolling, and there is a 30-day money-back guarantee if you move to a paid plan.

FAQs

Can AI Meeting Notes Replace My Own File Notes?

No. AI produces the draft and the verbatim transcript, you review, correct, filter, and sign off, because the adviser must ensure errors are caught before they enter the permanent record. The shift is from author to editor, not from adviser to spectator. For a practitioner-level discussion on where AI fits and where it does not, this interview with AdvisoryAI's CEO covers the distinction directly.

What If a Client Objects to Recording?

Document the refusal in your back office, revert to manual notes written immediately after the meeting, and have the client confirm the notes at the next meeting or in writing. Advisers should inform participants that recording may take place, so the consent conversation happens before the meeting starts.

How Do I Know Which Soft Facts Are Compliance-Relevant?

Ask whether the detail explains why the advice is suitable for this client: doubts, family influence, health concerns, life goals, and emotional reactions almost always qualify. Off-topic remarks and sensitive details unrelated to the advice should be filtered out during your review.

Do AI Meeting Notes Meet FCA Evidencing Standards?

They can, provided the output is structured, timestamped, source-traceable, and human-reviewed, because the FCA requires appropriate records and the obligation stays with your firm. A compliance check against Consumer Duty and COBS before the file is locked can add an additional layer of assurance.

How Long Should I Keep AI-Generated Meeting Transcripts?

Keep the finalised, signed-off file note in line with your standard record retention policy, and consider deleting raw audio promptly once the note is finalised under data protection principles. Retaining the transcript alongside the note can strengthen the audit trail where your compliance policy requires it.

Key Terms Glossary

Soft facts: Qualitative client details such as doubts, family dynamics, health concerns, and life goals that can evidence why advice is suitable.

Hard facts: Quantifiable fact-find data including age, income, assets, pension values, and ATR score.

COBS 9: The FCA's suitability rules, requiring reasonable steps to ensure recommendations suit the client's circumstances and a suitability report explaining why.

FG21/1: FCA finalised guidance on the fair treatment of vulnerable customers, requiring firms to understand the circumstances driving client decisions.

FG22/5: FCA finalised guidance on Consumer Duty, covering products and services, price and value, consumer understanding, and consumer support.

AI hallucination: A false or misleading output in AI-generated content, such as transcript text that misrepresents what was actually said.

Contemporaneous record: A record created at or near the time of the event, with timestamps showing when something was said or decided.