Marcus had been methodical. Six gaps closed in twelve days. AI policy written. Accountability mapped. Systems registered. Risks assessed. Data governance documented. Each answer to the Harcastle questionnaire had been hard-won, but they had been won.

He had two questions left — then he opened the wrong folder.

It was Jade's Q3 portfolio review for Harcastle Group. Sent three weeks ago. Before the ChatGPT suspension. He skimmed the executive summary, the asset allocation commentary, the ESG screening narrative. Well-structured. Confident. Professionally presented.

Then he searched the document for the word "AI."

Zero results.

He searched for "automated." Zero. "Generated." Zero. "Tool." Zero.

"Jade — this Q3 report for Harcastle. How much of this was ChatGPT?"

A pause on the Teams call. The kind that tells you the answer before the words arrive.

"Most of the narrative sections. I edited them. The numbers are mine. But the commentary drafts were all in ChatGPT. I thought that was just… how we used it."

Marcus looked back at the Harcastle questionnaire. Question 7.

Harcastle Group plc — AI Governance Due Diligence
Question 7 of 9
"Do you disclose to clients when AI systems have been used in the preparation of investment advice, portfolio commentary, or reports delivered to them?"
Please confirm your disclosure process, the form of disclosure provided, and whether clients are given the opportunity to object to AI-assisted work on their account.

He could answer nine of ten questions now. But question 7 was not about what he had built in the last twelve days. It was about what had already been sent.

⚠ Warning — FCA Enforcement Risk
Sending AI-assisted client communications without disclosure may constitute a breach of COBS 4.2 (fair, clear, and not misleading) and Consumer Duty PRIN 12 (acting in good faith). The FCA's AI & Technology guidance and Consumer Duty implementation reviews have both identified non-disclosure of AI use as a priority supervisory concern. Firms that cannot demonstrate a documented disclosure process are at elevated risk of regulatory intervention.

The Gap: What ISO 42001 Annex A.2 Requires

ISO/IEC 42001:2023 Annex A includes a set of controls that responsible AI system operators are expected to implement. Control A.2 — Responsible AI by Design — contains a disclosure requirement that most firms overlook because it does not appear in the main clause structure. It sits in the annex. But it is not optional.

The disclosure obligation under Annex A.2 has three components:

Component Requirement Obligation Level FCA Parallel
A.2.2 Affected parties shall be informed when AI has been used in processes that materially affect them SHALL Consumer Duty — Acting in Good Faith (PRIN 12)
A.2.3 The form and timing of disclosure shall be documented in the AI management system SHALL SYSC 3.2 — Record-keeping obligations
A.2.6 Where AI-generated content is presented to clients, the AI origin shall be identifiable or otherwise disclosed SHALL COBS 4.2 — Fair, clear, not misleading
A.6.1 Clients should be given opportunity to request human review of AI-assisted outputs that affect their interests SHOULD Consumer Duty — Products & Services / Price & Value

Whitmore & Associates had none of these. No disclosure language in their report templates. No documented process for informing clients when AI had been used. No record of who had decided the current approach was acceptable, because no one had ever consciously decided it.

ℹ Note — Materiality Threshold
ISO 42001 does not require disclosure of every AI tool used in a firm's operations. The obligation is triggered when AI output materially affects decisions made about a specific client. AI-drafted portfolio commentary, AI-generated risk assessments, and AI-produced suitability letters all meet the materiality threshold. AI used for spell-checking or internal summarisation typically does not.

Why This Matters Beyond Compliance

The disclosure requirement exists for a reason that goes beyond box-ticking. AI systems can introduce errors, biases, and confident-sounding inaccuracies. A client who does not know AI was involved cannot ask the right questions. They cannot request human review. They cannot apply appropriate scrutiny to commentary that may have been generated, not considered.

Jade's Q3 report for Harcastle contained ChatGPT-generated commentary on ESG screening criteria. That commentary referenced a regulatory standard — SFDR Article 8 classification thresholds — with complete confidence and marginal inaccuracy. The numbers were close, but not current. Jade had edited the prose but not checked the regulatory reference.

Harcastle had read it. Relied on it. Made no complaint, because they had no reason to know the source.

⚡ Caution — Retroactive Exposure
Where AI-assisted content has already been sent to clients without disclosure, firms face potential liability even before any complaint is raised. Under Consumer Duty, the obligation to act in good faith is prospective and ongoing. Firms should consider whether a client communication programme is required for material prior AI use — and take legal advice before deciding not to disclose.

The Before and After: What a Disclosure Looks Like

Marcus spent forty-five minutes on Day 13 pulling every client-facing document template Whitmore used: the quarterly portfolio review, the annual suitability report, the rebalancing recommendation letter, and the ESG screening summary. None of them contained disclosure language.

He built a disclosure clause. Then he compared what Harcastle had received against what they would receive from Day 14 onward.

✕   Before — Q3 Report (Sent Without Disclosure)
✓   After — Compliant Template (DOC-AIMS-FIN-DISC-001)
WHITMORE & ASSOCIATES — Q3 PORTFOLIO REVIEW

Executive Summary

Your portfolio delivered a return of 3.2% in Q3, outperforming the benchmark by 0.8%. The ESG screening applied across your holdings continues to align with SFDR Article 8 classification thresholds, with exposure to excluded sectors at 0.4% — within tolerance…


[No AI disclosure. No reference to tools used. No client rights statement.]


Prepared by: Jade Nwosu, Relationship Manager
Reviewed by: [blank]
Date: 14 September 2024

WHITMORE & ASSOCIATES — Q3 PORTFOLIO REVIEW (REVISED)

Executive Summary

Your portfolio delivered a return of 3.2% in Q3, outperforming the benchmark by 0.8%. The ESG screening applied across your holdings continues to align with SFDR Article 8 classification thresholds, with exposure to excluded sectors at 0.4% — within tolerance…

AI Assistance Disclosure

Sections of this report were drafted with the assistance of AI language tools (Microsoft Copilot, AIMS-SYS-002). All AI-assisted content was reviewed and approved by the named Relationship Manager prior to issue. If you would like any section reviewed by a qualified adviser without AI assistance, please contact us directly.

Authorised by: Marcus Osei, Head of Compliance (SMF16)
Left: The Q3 report sent to Harcastle Group. Right: The revised template from Day 14. The disclosure block is mandatory on all AI-assisted client documents under DOC-AIMS-FIN-DISC-001.
⚙ ITIL 4 — Continual Improvement
Template governance is a service management problem, not just a compliance problem. Disclosure clauses embedded in document templates are only effective if template version control is maintained, old templates are retired, and staff are trained on which version to use. ITIL 4's Document and Knowledge Management practice provides the operational framework for this — ensuring the right template reaches the right person at the right time.

Building DOC-AIMS-FIN-DISC-001: The Client Disclosure Framework

Marcus did not just update the Q3 template. He built a disclosure framework that could be applied consistently across all client-facing outputs, regardless of which AI system was used or which member of staff had prepared the document.

DOC-AIMS-FIN-DISC-001
AI Client Disclosure Framework — Whitmore & Associates Ltd
Annex A.2 / A.2.2 / A.2.3 / A.2.6
Consumer Duty PRIN 12 / COBS 4.2
Marcus Osei, SMF16
Quarterly — triggered on new AI system onboarding
All client-facing written communications where AI has materially contributed to content
Any AI-generated content comprising >20% of a client document section

The framework defined three elements: the disclosure template language, the threshold at which disclosure is required, and the sign-off process. Each is documented below as it was embedded into Whitmore's templates.

DOC-AIMS-FIN-DISC-001 — Standard Disclosure Clause (Version 1.0)
AI Assistance Disclosure: [Section name(s)] of this document were drafted with the assistance of [AI System Name — AIMS-SYS-00X]. All AI-assisted content was reviewed and verified by [Named Adviser] prior to issue. You may request a version prepared without AI assistance by contacting your Relationship Manager.
Automated Decision Notice: The [assessment type] referenced in this communication was produced using an automated AI system ([AIMS-SYS-00X]). A qualified member of the compliance team has reviewed this output. If this assessment affects a decision about your account, you have the right to request human review under Article 22 of UK GDPR.
Disclosure is required when AI-generated content constitutes >20% of any named section in a client document, OR where AI has contributed to any recommendation, suitability determination, or risk assessment affecting the client, regardless of the percentage threshold.

What Harcastle's Question 7 Actually Asked

Question 7 had three sub-parts. Marcus addressed each one.

Sub-question Marcus's Response Evidence Document
Do you disclose AI use to clients? Yes — from Day 14, all client documents use the disclosure clause in DOC-AIMS-FIN-DISC-001. Prior reports are under review. DOC-AIMS-FIN-DISC-001 v1.0
What form does disclosure take? A standardised disclosure block embedded in the document footer section, identifying the AI system by AIMS-SYS reference, the sections affected, and the human reviewer. Template library v1.0
Can clients object to AI-assisted work? Yes — the disclosure clause includes a client opt-out provision. Clients may request a non-AI version of any document. Response SLA: 5 business days. PROC-AIMS-FIN-HITL-001 (forthcoming)

The third sub-answer referenced a document that did not yet exist: PROC-AIMS-FIN-HITL-001, the Human-in-the-Loop oversight procedure. Marcus had written a placeholder. He had eight days left and two gaps to close.

F-G7
✅ Provisionally Satisfied
DOC-AIMS-FIN-DISC-001 v1.0
Retroactive review of AI-assisted documents sent in prior 8 months
7 of 21

Marcus sent the updated framework to Diana at 11:42pm on Day 13. No preamble. Just the document.

She replied at 7:08am.

"The retroactive review worries me more than the framework. If Harcastle ask whether we disclosed on their Q3 report, we cannot say yes. We need to be prepared for that conversation."

Marcus had already drafted the note. A brief, factual letter to Harcastle acknowledging that their Q3 review had included AI-assisted content, confirming that no material inaccuracies had been identified upon human review, and confirming that all future reports would carry full disclosure.

He had not sent it. He wanted legal sign-off first. But the draft existed.

One problem at a time. Two gaps left. Seven days on the clock.

"Gap 8," he typed to himself. "Human oversight. What happens when the AI is wrong — and no one was watching."
ℹ Note — Regulatory Trajectory
The FCA's February 2025 AI update and the EU AI Act (applicable to UK firms with EU clients from August 2026) both reinforce AI transparency obligations. The direction of travel is clear: disclosure of AI use in client-facing financial communications is becoming a baseline expectation, not a best-practice aspiration. Firms that establish disclosure frameworks now are better positioned for both regulatory reviews and client due diligence requests.
Quality Record — TMPL-AIMS-ART-001 Gates
✓
Gate 1: ISO clause correctly cited (Annex A.2)
✓
Gate 2: FCA regulatory parallel mapped (Consumer Duty / COBS 4.2)
✓
Gate 3: Fictional scenario consistent with REF-AIMS-FIN-SCEN-001
✓
Gate 4: Fix document produced (DOC-AIMS-FIN-DISC-001)
✓
Gate 5: Unique visual element (before/after disclosure comparison)
✓
Gate 6: Notice hierarchy maintained (WARNING, CAUTION, NOTE, ITIL)
✓
Gate 7: Story continuity from F-G6 (day counter advanced to 12–14)
✓
Gate 8: No DANGER notices used
✓
Gate 9: Diana's voice used sparingly (one response)
✓
Gate 10: Cliffhanger leads to F-G8 (HITL placeholder cited)
Next in the Series
DOC-AIMS-FIN-DISC-001 told clients that a human had reviewed the AI output. But what did that review actually look like? Was there a process? A checklist? Any record that the review had happened at all?
Marcus had promised Harcastle a human oversight procedure. He had seven days to build one — and a growing sense that "human review" at Whitmore had meant reading it once before hitting send.
→ F-G8: Human-in-the-Loop Oversight — Coming Next