Aipedia: Insurance brokers

AI pricing and underwriting under delegated authority

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

A broker or MGA pricing risks with an algorithm under a binding authority is responsible for how that algorithm treats customers, as well as for staying within the authority. Bias, data errors or breaches of the authority can lead to claims from both clients and insurers.

Part of: AI in insurance brokers

In short

Under delegated authority, brokers and MGAs quote and bind on an insurer's behalf. Many now use algorithmic pricing or rating engines to do so. The FCA AI Update lists the Equality Act 2010 and UK GDPR Article 22 under fairness, and the Data (Use and Access) Act 2025 added new rules on solely automated decisions from February 2026. If the engine binds risks outside the authority, misprices, or produces unfair outcomes, the insurer may seek recovery from the broker and customers may complain. Which policy responds, broker E&O or something else, depends on the facts and the wording.

How brokers and MGAs use algorithmic pricing

Under a binding authority, an insurer allows a broker or MGA to accept risks and set premiums within agreed rules. Increasingly, those rules are run by a rating engine or algorithmic underwriting model that scores risks, applies loadings and declines or refers cases automatically. Some use machine learning trained on claims data or third-party data to price more finely.

The 2024 Bank of England and FCA survey found 55% of AI use cases involved some automated decision-making, and pricing and underwriting are obvious examples.

The duties that apply

Fairness: the FCA AI Update lists the Equality Act 2010, UK GDPR Article 22, PS22/9 and FG22/5 under fairness and contestability. Pricing that disadvantages people because of a protected characteristic, directly or through a proxy, creates legal and regulatory risk.
Automated decisions: section 80 of the Data (Use and Access) Act 2025 inserted Articles 22A to 22D into UK GDPR, fully in force from 5 February 2026, with safeguards for solely automated significant decisions and restrictions on special category data such as health data.
Consumer Duty: the price and value outcome applies to retail products, and PRIN 2A.2.8R requires firms to avoid foreseeable harm.
SM&CR: the FCA expects a named senior manager to own each AI use case.

What can go wrong

Illustrative: an MGA's rating engine is updated with a new data feed. A mapping error means a class of trades the binding authority excludes is rated as an acceptable trade, and the engine binds dozens of policies outside the authority over several months. When a large claim arrives on one of them, the insurer says the risk was never within the authority and seeks to recover from the MGA.

A different problem arises if the model uses a variable, such as a postcode or a data field linked to disability, that operates as a proxy for a protected characteristic. That can lead to complaints, regulatory scrutiny and insurer audit findings, even where every policy is within the authority.

Insurers commonly audit delegated authority business. An audit that finds poorly governed models, missing change logs or unexplained pricing outcomes can lead to tighter terms, loss of the authority or a claim, whatever the eventual loss experience.

Broker E&O versus insurer-side liability

The line matters. A customer harmed by unsuitable advice or a misstatement usually claims against the broker or MGA, and broker E&O cover would commonly respond. Where the broker or MGA breaches the binding authority, the insurer itself may claim against it, and whether that claim falls within the E&O policy depends on the wording, as some policies treat claims by insurers or binding authority breaches differently. Regulatory fines are often excluded. For how roles split, see MGA PI insurance and delegated authority.

Practical controls

Hard-code the binding authority limits into the engine and test them after every change.
Reconcile bound risks against the authority monthly and report exceptions to the insurer.
Test rating outcomes for proxies of protected characteristics and document the results.
Keep a log of model versions, data sources and who approved each change.
Provide a route for customers to ask for human review of an automated decision.
Prepare for insurer audits with clear evidence of model governance.

How Apex helps

Apex arranges PI for insurance brokers and MGAs. We can review how your E&O policy treats claims brought by capacity providers as well as by clients, and how insurers view algorithmic pricing.

Sources

Frequently asked

Who is responsible for bias in a delegated authority pricing model?

Both the insurer and the broker or MGA can carry responsibility, depending on who designed and runs the model and what the agreement says. The FCA expects each firm to own its AI use cases, and outsourcing the model does not remove responsibility.

Does UK GDPR restrict automated pricing decisions?

The Data (Use and Access) Act 2025 replaced the old Article 22 with Articles 22A to 22D, in force from 5 February 2026. They require safeguards for solely automated significant decisions and restrict the use of special category data such as health information.

Will broker E&O cover an insurer's claim for a binding authority breach?

It depends on the wording. Some E&O policies cover claims by insurers arising from breach of authority, others restrict or exclude them. Brokers and MGAs with delegated authority should check this specifically with their own broker.

Should customers be able to challenge an algorithmic price?

Where a decision is solely automated and significant, the new UK GDPR provisions require safeguards, including the ability to contest it and obtain human intervention. Offering a clear review route is also consistent with Consumer Duty expectations.

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.