FCA authorised · FRN 724952 0117 325 0027Quote & buy →
Aipedia: Insolvency practitioners
The Insolvency Code of Ethics requires an IP to decide whether technology output is appropriate for its intended purpose before using it (R321.1). With no RPB AI guidance, these paragraphs are the main rules an IP's use of AI will be judged against.
Part of: AI in Insolvency practitioners
In short
The joint Insolvency Code of Ethics, approved by the JIC and incorporated by the IPA, contains technology provisions that apply to AI. R321.1 requires an IP to determine whether using technology output is appropriate for the intended purpose. Application material flags threats from insufficient data or expertise (115.1 A6), the role of oversight and training (116.2 A3) and automation bias (117.3 A2). Supervision (R1.10, R103.2), confidentiality (R104.3 to R104.6) and records (R130.2) duties also apply. No RPB has issued AI-specific guidance.
Insolvency practitioners are authorised and regulated by the recognised professional bodies (RPBs), including the IPA and ICAEW, under section 390 of the Insolvency Act 1986. The Insolvency Service oversees the RPBs and publishes current sanctions they report in an agreed format, applying the Common Sanctions Guidance. The Joint Insolvency Committee (JIC) approves the joint Insolvency Code of Ethics, which the IPA incorporates into its own Code of Ethics for members.
None of these bodies had published AI-specific guidance as at October 2026. Checks of the IPA's regulation guidance list, its 2026 regulatory updates, R3's policy pages and Insolvency Service collections found nothing on AI. R3's August 2026 report confirms the gap: practitioners describe "the lack of clear, practical guidance from the Recognised Professional Bodies" on permissible AI use.
So the practical rulebook for AI is the Code itself. It does not mention AI by name, but its technology paragraphs, which mirror the IESBA technology revisions adopted by the accountancy bodies, are broad enough to cover tools like this.
The references below are from the current IPA Code of Ethics PDF (current copy, uploaded June 2026).
| Paragraph | What it says | Status |
|---|---|---|
| R321.1 | When an IP "intends to use the output of technology", they "shall determine whether the use is appropriate for the intended purpose". | Requirement (binding) |
| 115.1 A6 | Gives examples of "facts and circumstances relating to the use of technology" that might create threats, such as insufficient data or a lack of expertise. | Application material |
| 116.2 A3 | Evaluating technology threats depends on factors such as corporate oversight and staff training. | Application material |
| 117.3 A2 | Lists examples of bias, including "automation bias", a tendency to favour output from automated systems. | Application material |
Paragraphs with an "R" prefix are requirements. "A" paragraphs explain how to apply them, and an RPB would expect members to take them into account. Read together, they say an IP may use AI but must judge its fitness for each task, recognise when data or expertise is too thin for the output to be trusted, and actively guard against the instinct to accept what the machine produces.
Much AI use in insolvency practice may fall to case administrators and managers rather than the appointment holder. The Code makes the IP answerable for that work:
R320.5 is a useful hook for third-party AI platforms used for bank analysis, document review or creditor portals: the arrangement should be reviewed, not just set up once. R3's informal-use finding, that generative AI is "mostly used individually and informally", sits uneasily with R103.2, because an IP cannot supervise use they do not know about.
The confidentiality duties at R104.3 and R104.4 (prohibited uses and disclosures), R104.6 (a continuing duty) and R316.1 apply to any data put into an AI tool. Pasting debtor or creditor information into a consumer chatbot that may retain or reuse it is the obvious risk; see debtor and creditor data in AI tools.
The records provisions matter just as much. Under 130.1 A1 an IP should be able to justify actions through "written contemporaneous records", and under R130.2 and 130.2 A1 those records should let a reasonable and informed third party assess the actions taken. If an AI tool shaped a decision, a file that records only the outcome may not meet that standard. Conflicts and ethical issues must also be documented (R312.9 and R140.2).
ICAEW-licensed IPs should also note the ICAEW Code technology revisions in force from 1 July 2025; see the accountants guidance explainer. R3 has recommended that RPBs publish practical guidance on data security, human oversight, quality assurance and disclosure, and some interviewees suggested a dedicated SIP on AI. Neither existed at the time of writing.
For a working practice, the Code translates into a short list: approve tools centrally, train staff to challenge output, make a named person responsible for checking each AI-assisted deliverable, and record that check on the case file. These steps also help if a claim arrives. A defensible file showing proper review is often the difference between a claim that can be defended and one that has to be settled.
The Code does not change what PI covers. Compulsory PI under the RPB rules, such as the IPA PII guidance, says nothing about AI, and R3 found some practitioners unsure whether errors from AI use are covered. A breach of R321.1 is primarily a regulatory matter; a creditor's negligence claim arising from the same error is what PI usually responds to, subject to the wording. Regulatory defence costs cover varies between policies. See bond vs PI for AI errors, AI for SIP 2 investigations and PI for insolvency practitioners.
Apex arranges PI for insolvency practitioners, and a broker can check how your wording treats AI-assisted work and regulatory investigations. Return to the insolvency practitioners Aipedia hub.
It does not use the term AI, but its technology provisions apply to it. R321.1 requires an IP intending to use technology output to determine whether the use is appropriate for the intended purpose, and 117.3 A2 gives automation bias, a tendency to favour output from automated systems, as an example of bias.
Yes. Paragraphs with an R prefix are requirements of the Code, which the JIC approves and the IPA incorporates for its members. The A paragraphs are application material explaining how to comply, and RPBs would expect them to be taken into account.
No RPB or Insolvency Service AI-specific guidance was found as at October 2026. R3 has called for practical guidance on data security, human oversight, quality assurance and disclosure, and some practitioners have suggested a dedicated SIP on AI.
The IP. R1.10 says work done by the insolvency team on the IP's behalf must comply with the Code, and R103.2 requires the IP to ensure that people working under their authority are appropriately trained and supervised.
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.
