We built a tool that takes over much of the grind in an insurance broker’s day, and we showed it at this year’s Forsikringsmeglernes dag, the Norwegian insurancebroking industry’s main event, in front of more than 500 brokers. Two of our people put it together in a matter of weeks. The room had just heard from BCG that commercial broking has lagged on technology for years.
Take tenders, the task brokers name first when you ask where their time goes. Most of a tender document is reuse and standard structure, so the tool assembles the package from a few inputs and leaves the broker to handle the parts that need a person. Ask it a question and it answers across policy wordings, regulation, and your own internal notes at once. It runs prospecting, renewals, and the rest of the pipeline in the same place. What it doesn’t touch is the judgment and the advice a client pays for. That stays the broker’s, by design, and the tool only clears the repetitive work stacked in front of it. The brokers watching hadn’t seen anything built for their work that did this.
None of us had worked as insurance brokers before, and that helped more than it hurt. An expert knows every reason things are done the way they are; we don’t, so we ask the questions that sound stupid and build past the constraints an insider treats as permanent. Prospecting shows it. The version we built takes a company number and returns a quick risk score, and an insider would stop there, because that’s what the data cleanly supports today. We kept asking what else a machine could watch for a broker: a company winning a contract that changes its exposure, or a new risk manager turning up in its hiring feed. None of that is exotic. It just takes someone without a fixed idea of where the tool is supposed to stop.
The approach is the same wherever we point it. Sit down with the people who do the work, find the task that quietly eats their week, and build something that takes it off their desk. Insurance broking is one industry. The bottleneck in yours is different, and we’d have to learn it, but the way in doesn’t change.
A few years ago, a tool like this would have been a year-long project with a real budget and a specialist team behind it. The models and the tooling have improved enough that a small, capable team can now do in weeks what used to take all of that. The tools don’t wield themselves, though. Knowing what to build, and building it well and fast, is its own skill, and it’s the one we’ve spent years on. We can put a working prototype in front of you, the kind your people can click through and react to, in a fraction of the time you’d expect.
Markets that change this fast reward companies willing to break a few of their own rules. Most firms haven’t built something like this, and the reasons are ordinary: they don’t know how, they can’t spare the people, or those people are buried in everything else. So it stays on the list of things to do someday. We do this for a living. Talk to us before your competitor does.


.png?width=354&height=354&name=Bridging%20your%20goals%20with%20trusted%20AI%20(2).png)