Skip to main content
Insurance

Submissions, schedules and bordereaux, read and checked for you. The risk decision stays with your underwriters.

An underwriting team's day fills up with documents that must be opened, read, keyed and checked. A submission arrives as a PDF and a spreadsheet. Bordereaux arrive as spreadsheets, each coverholder laying theirs out differently and changing the layout without warning. The work demands real skill where it matters and is mechanical everywhere else.

The cost is in the handling, not the decision

Three numbers show how well an underwriting team is running: the share of submissions that receive a quote, how long brokers wait for it, and how much premium and commission slips through the bordereaux because nobody has time to reconcile every line.

All three come down to the same constraint. Someone has to open the PDF, read the schedule, compare it with the last one and key it into the system, and the only people who can do that reliably are the people you would rather have pricing risk.

Led by an engineer who has built systems inside an MGA

Fusion AI Consulting's founder, Jack Quinn, spent his career building production systems in regulated finance, most recently at CFC, the largest independent MGA in the UK, with its own Lloyd's syndicate. The systems he worked on there processed over £3bn a year.

At CFC he built a risk rules engine, a service that turned inbound email into structured data, a firm identification service that worked out which company a submission referred to, and a bordereaux enrichment service. They correspond directly to four of the problems on this page: appetite and triage, submission intake, matching what arrives against what you already hold, and getting coverholder returns to reconcile.

All four ran in production, at live volumes, inside a regulated business. None of them was a pilot.

A submission, before and after

Illustrative. Appetite, rating and the decision to quote stay with the underwriter.

What we build for insurers and MGAs

These are the problems insurers and MGAs raise with us most often. It is not an exhaustive list: discovery decides what your firm needs, and occasionally the answer is none of the below.

Not your sector?AccountancyPrivate Equity and Banking

Shadow mode

Proved on your own work before it goes live

No system takes over any part of your operation on the day it is finished. It runs in shadow mode first, alongside the people doing the work today. It reads the same submissions, triages the same inbox and reconciles the same bordereaux, while every reply to a broker and every entry in your systems stays with your team. Your existing process carries on exactly as before, so at no point are you relying on something unproven.

You can then set the order it would have ranked the submissions in, the figures it read and the discrepancies it raised against what your underwriters found. Shadow running lasts as long as you want it to, on your own work.

When you are satisfied with the results, you decide when it goes live, and you can start with part of the work. From then on, the system does the work it has shown it can do.

Questions

The first conversation is free. Beyond that, we quote a fixed price once we both know exactly what is being built, and nothing is charged until the scope and the price have been agreed in writing.

Discovery usually takes one to two weeks. Build time depends on the system, and the timetable is fixed in the quote.

No. We handle discovery, design, build, deployment and testing. Your team's role is to explain how the work is done today and to approve what we propose.

It does not go live until that question has been answered with evidence. Every new system first runs in shadow mode: it works on your real cases alongside your team, but changes nothing and contacts no one. You compare its output with your team’s for as long as you need to, and it goes live only when you are satisfied and decide to switch it on. Once live, anything it is unsure about is passed to a person.

Your data stays in your own environment wherever the work allows. Access is limited to what each system needs, every action is logged, and a person approves anything that requires judgement. Data is handled in line with UK GDPR and is never used to train AI models. The same rules apply to any broker or coverholder data involved.

The policy administration, document management and email systems you already use, and the spreadsheets that sit between them, including each coverholder’s own bordereaux layout. Compatibility is confirmed during discovery, before you commit to anything.

No. Every system sits in your own cloud account, is built on standard, well-documented technology that any competent developer can maintain, and is handed over with full documentation. You can keep us on for support, but you never have to.

Start with a conversation

The first conversation is free. We will talk through where your team’s time is going and tell you plainly whether we can help.

Fusion AI Consulting

Fusion AI Consulting designs, builds and runs bespoke systems that give accountancy, insurance and investment firms more capacity without adding headcount. Based in London, working across the UK.

Contact
© 2026 Fusion AI Consulting Ltd. All rights reserved.Fusion AI Consulting Ltd is registered in England and Wales, company number 16771023. Registered office: 124 City Road, London, EC1V 2NX.