Aipedia: Insolvency practitioners

AI and estimated outcome statements

Aipedia, by the Apex broking team · Last reviewed 2026-10-27

AI can build statements of affairs and estimated outcome statements, but the IP must test the assumptions because creditors make decisions on the figures. Overstated recoveries can lead to fee challenges, conduct complaints and claims.

Part of: AI in Insolvency practitioners

In short

Estimated outcome statements drive creditor decisions on proposals, fees and write-offs. AI tools can produce them quickly from the books, but generic recovery rates, missed secured or preferential claims and understated costs can overstate returns. The Insolvency Code of Ethics requires the IP to decide whether output is appropriate (R321.1), warns of insufficient data and automation bias, and expects assumptions to be recorded (R130.2). Fee challenges and complaints may follow; a creditor who relied on the figures may claim, and PI usually responds, subject to the wording.

How AI is used to build outcome estimates

Statements of affairs and estimated outcome statements tell creditors what is likely to be recovered and who will be paid. They inform decisions on whether to support a proposal, approve fees, fund litigation or write off a debt. AI tools can now assemble them quickly: extracting balances from the books, applying recovery percentages to asset classes, estimating costs and producing a waterfall of returns to each class of creditor.

The output looks authoritative because it is precise. Precision is not the same as accuracy, and an estimate is only as good as the realisation assumptions behind it.

The duties that apply

Under the Insolvency Code of Ethics:

R321.1: the IP must determine whether using the tool's output is appropriate for the intended purpose, here a figure creditors will act on.
115.1 A6: insufficient data. Book values in a failed company rarely reflect what assets will fetch.
117.3 A2: automation bias, including accepting default recovery rates without challenge.
R320.3: where the estimate relies on agents' or valuers' figures, evaluate whether their work is appropriate for the purpose.
130.1 A1 and R130.2: record the assumptions so a reasonable and informed third party could assess them.

An estimate that drives a creditor vote also engages objectivity. If the figures happen to favour the procedure the IP proposed, they will be scrutinised closely.

What can go wrong

Illustrative scenario: in an administration, an AI tool builds the estimated outcome statement from the management accounts. It applies a standard recovery rate to book debts, not knowing that most of the ledger relates to one customer disputing the invoices. The statement shows a meaningful return to unsecured creditors, who approve the administrator's proposals and fee basis on that footing. The debts prove largely irrecoverable. Creditors challenge the fees and complain that they were misled into supporting the proposals.

Typical failures:

book values or generic recovery rates applied to assets with specific problems;
costs of realisation, including the IP's own fees, understated;
secured or preferential claims missed, inflating the unsecured return;
figures not updated in later progress reports as the picture changes.

A tool that invents a plausible figure where data is missing is a particular hazard; see AI hallucination liability.

Fee challenges, conduct and PI exposure

Overstated recoveries can lead to three kinds of consequence. Creditors may challenge fees, which is usually a matter for the procedure and the court rather than insurance. They may complain to the RPB about conduct; regulatory defence cost cover depends on the wording. And a creditor that lost money by relying on the estimate may bring a negligence or misrepresentation claim.

PI usually responds to that third category, subject to the policy terms, the excess and any exclusions. The IPA's PII guidance makes PII compulsory for individual members, and ICAEW's PII Regulations apply to insolvency bodies it regulates. Neither addresses AI, and R3 found some practitioners unsure how their policies would respond to AI errors. Ask your broker rather than assume.

Practical controls

Replace default recovery rates with case-specific assumptions, and write down the reason for each.
Show ranges or sensitivities where outcomes are uncertain.
Reconcile secured, preferential and unsecured claims to the claims register.
Use agents' or valuers' figures for material assets; see asset tracing and realisations.
Have someone other than the preparer challenge the statement before it goes to creditors.
Carry the same assumptions into pre-appointment advice so the story is consistent.

How Apex helps

Apex arranges PI for insolvency practitioners. A broker can review how your policy treats creditor claims arising from estimates and whether regulatory costs are included. See the PI guide for IPs and the insolvency practitioners hub.

Sources

Frequently asked

Can AI prepare an estimated outcome statement?

It can assemble one quickly, but the IP must decide under R321.1 whether the output is appropriate for a figure creditors will act on. In practice that means replacing generic assumptions with case-specific ones and recording why.

What happens if an estimate overstated recoveries?

Creditors may challenge fees, complain to the RPB or, if they relied on the figures and lost money, bring a claim. PI usually responds to negligence claims, subject to the wording, while fee challenges are typically dealt with under the insolvency procedure itself.

Should estimates show ranges rather than single figures?

Where outcomes are genuinely uncertain, ranges or sensitivities give creditors a fairer picture and make the IP's judgement clearer. A single precise figure produced by a tool can suggest more certainty than exists.

Check your PI covers how you use AI

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 callback

Apex 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.