One of the biggest platform shifts in engineering over the past few years has been that AI models can now write code. Work that used to take a team of engineers several weeks now spins up in an afternoon for a single developer, and that change has rewritten who gets to create software inside a company.

For the past year we've been asking a simple question: what is the equivalent shift for insurance? What happens when code generation isn't only an engineering capability, but something underwriters, claims adjusters, broker partners, and program managers can use directly, inside the tools they already work in every day?

Today we're launching Code Gen inside the FurtherAI assistant, and the answer to that question is starting to come into view.

The artifacts insurance teams wait weeks for now take minutes

Here's what this looks like in practice.

An underwriter drops a submission into the FurtherAI assistant and asks for a deck, and the assistant builds a complete submission profile (with risk summary, exposure breakdown, loss history, and pricing rationale) formatted the way the team presents to committee. The whole thing is done in minutes.

That same underwriter can point the assistant at their portfolio and ask for a dashboard showing concentration by class of business, geography, limit bands, and loss ratio trends. The dashboard simply builds inside the assistant and updates as the underlying data does.

Program managers, on the other hand, can spin up microsites for a new program that covers appetite, submission guidelines, and a contact flow. With Code Gen, complete sites are ready to share with producing brokers by the end of the day.

Insurance expertise stays with the operator

Until now, an underwriter who wanted a custom dashboard had two real options: wait for engineering or BI to build it, which meant filing a ticket and accepting a quarter of patience while a roadmap conversation happened, or build something in Excel that approximated the answer and accept that it would never reach the rest of the team in a useful form.

Code Gen removes that tradeoff entirely, and the person who understands the work most deeply — the underwriter writing the policy or the adjuster handling the claim — can now produce the artifact that moves the work forward.

Insurance runs on artifacts, and that's why this matters

Every meaningful workflow in insurance produces an artifact that someone else relies on to make a decision, and the list runs across the entire industry: committee decks, portfolio reviews, loss summaries, renewal analyses, producer-facing one-pagers, reinsurance treaty summaries, and more. When the production of those artifacts takes days, the decision-making rhythm of the business slows to match. Quarterly portfolio reviews stretch into six-week production cycles, renewal decks land too late to drive a real conversation with the broker, and producer one-pagers go out the door already stale by the time they arrive in front of an agent.

Code Gen collapses that production time. An underwriter builds the committee deck during the conversation it's meant to inform, a portfolio lead drafts the review inside the leadership meeting where it's needed, and a program manager spins up the producer one-pager during the broker call itself. The rhythm of the business finally catches up to the rhythm of the work.

Bring Code Gen to your underwriting and claims teams 

Code Gen is available to FurtherAI customers starting today, and we're rolling it out across underwriting, claims, and program management teams. 

If you want to see Code Gen running against your own submissions or book, get in touch.