FCA authorised · FRN 724952 0117 325 0027Quote & buy →
Aipedia: Management consultants
Certified Management Consultants must now tell clients when AI materially shapes their work, and the ICMCI AI Code expects engagement terms to cover AI scope, data, IP and restrictions. Clear terms reduce disputes and help with insurers.
Part of: AI in Management consultants
In short
The ICMCI Code of Responsible Use of AI, launched in June 2026, requires CMCs to tell clients when AI materially shapes analysis, recommendations or deliverables (3.3, 7.1). Uses involving personal data, profiling or automated decisions generally need explicit client agreement (7.2), and engagement terms should cover AI scope, IP, data protection and restrictions (7.3). Some clients, including public bodies under PPN 017, set their own AI rules. Written terms help defend a claim by showing what the client agreed, and they make it easier to describe your AI use accurately to your PI insurer.
Engagement letters may say nothing about the tools a consultant uses, and generative AI makes that gap matter. Clients may worry about their data being used to train public models, about paying consultant rates for machine-generated work and about reports containing fabricated material, a concern sharpened by the Deloitte refund in Australia reported by CFO Dive. Some clients may ban AI outright; others allow it with conditions.
Without written terms, a dispute about AI use becomes a dispute about what the client understood, which is hard to win after the event.
The ICMCI AI Code is mandatory for individual CMCs. Its engagement provisions are:
| Clause | Requirement | What to put in writing |
|---|---|---|
| 3.3 / 7.1 | Tell clients when AI materially shapes analysis, recommendations or deliverables | A statement of the tasks for which AI may be used, and a note in deliverables where it materially shaped the work |
| 7.2 | Uses involving personal data, profiling or automated decisions generally need explicit client agreement | Specific consent for those uses, with the data involved and the safeguards |
| 7.3 | Terms should cover AI scope, IP, data protection and restrictions | Approved tools or categories, ownership of outputs, data handling, and any client prohibitions |
| 3.4 / 8.3 | No confidential data in tools that may reuse or train on it without safeguards | A commitment on tool settings and data retention |
| 8.4 | Check outputs for IP infringement | Who owns deliverables and how third-party material is cleared |
Some clients prohibit AI use in their engagements, or restrict it to tools they approve. Breaching such a term is a breach of contract even if the work is accurate. Public bodies may ask about AI use during procurement under PPN 017, and contract terms may bar using buyer data to train AI without written approval. If you answered an AI question in a bid, your engagement terms and practice should match that answer. See AI bid writing and PPN 017.
Illustrative: a consultant uses AI tools to draft parts of a review without mentioning it. The client later learns this from metadata in a document and, unhappy with the findings, argues it was misled about the nature of the work, refuses to pay the final invoice and claims the report is unreliable. The engagement letter says nothing about AI either way. The consultant must now defend both the quality of the work and its own candour.
With a clear clause and a short disclosure in the report, the argument would have been confined to whether the work was done with reasonable skill and care.
Related: the ICMCI AI Code explained and AI benchmarking with client data.
Apex arranges PI for management consultants and can check how AI clauses in your engagement terms interact with your policy. See the AI in management consulting overview.
If you are a Certified Management Consultant, ICMCI 7.3 expects engagement terms to cover AI scope, IP, data protection and restrictions, and 3.3 requires disclosure where AI materially shapes the work. For others it is not mandatory but is good practice.
Then using it would breach the contract, even if the work is accurate. Agree in writing what the ban covers, for example whether spelling checkers or transcription count, and make sure your team knows the restriction applies to that engagement.
It can. Clauses that warrant AI outputs or give broad indemnities may create contractual liability beyond what PI covers. Clauses that describe AI use and confirm human review generally help. Show unusual terms to your broker before signing.
Tell us about your firm and the work you do. A broker will review how your current PI treats AI-assisted work, or arrange new cover. No fee for a quote. Or call 0117 325 0027.
Start a PI quote Request a callbackApex Insurance Brokers Limited is authorised and regulated by the Financial Conduct Authority. Registered in England and Wales, company number 07014570. This page is general information about how AI is used in the profession and how insurance may respond. It is not legal, regulatory or insurance advice on your circumstances, and it does not guarantee that cover will be available or on what terms.
