For RC-data API companies, add each insurer's vehicle code to your response

If you sell RC lookups, a good share of your customers are in motor insurance. Brokers, gateways, agents' apps and insurers call your API to fill in the vehicle so that a customer doesn't have to type it. And sooner or later one of them asks the obvious next question: "Can you also give us the insurer's code for this vehicle?"

It's a reasonable request. They're calling you for the vehicle so that they can quote it, and the RC alone doesn't get them a quote. Every insurer wants its own make, model and variant code, and turning CRETA 1.6 VTVT SX into five insurers' codes is the step they still do themselves.

Why it's harder than it looks

On the face of it you already have the data, so it's just a lookup. In practice it means taking on a new kind of work.

You'd need each insurer's vehicle master and the habit of refreshing it as models launch. You'd need a way to match RC spellings, which are messy, to master listings, which are messy in a different way. And you'd need to be right often enough to put your name on it, because your customers will hold you to it the first time a quote goes out with the wrong variant.

Most RC-data teams are good at what they do: reliable, fast, well-documented lookups. Vehicle mapping is a separate problem with its own long tail, mostly two-wheeler variants and models that share names. Doing it well is a project in its own right, not a sprint at the end of a quarter.

What Dorvy Tech adds

Dorvy Tech takes the vehicle fields you already return (maker and model as written, fuel, cc, seats, body type, norms and class) and gives back, for each insurer master, the listing and code, a confidence level of HIGH, MEDIUM or LOW, the closest alternatives, and a plain "no listing" with a reason when the insurer doesn't cover that vehicle.

It's built to sit inside your product:

• It takes about 7 ms per vehicle for five insurers, which is small next to an RC lookup.
• It runs on your own servers on an ordinary CPU. It needs no GPU and makes no outbound calls, so your customers' data stays where it is today.
• It needs no registration numbers or owner details, only the vehicle fields.
• It goes out under your own brand. Your customers see your API and your field names.
• We set up each new insurer master and test it before it goes live for you. Mapping to a customer's own vehicle master is available as a setup project.

The commercial case

You'd be selling a better answer to a question your customers already ask, to people who already pay you per lookup. An RC response with insurer codes is worth more to a motor insurance customer than one without, and it's harder for a competitor to copy than a cheaper RC call.

The accuracy is what makes it sellable. On our held-out test of 150 real RC records across five insurer masters, Dorvy Tech picked the exact listing 77.0% of the time, against 66.2% for the best standard matcher, and made less than half the wrong picks, about 13% against 32%. The test set was deliberately hard, and how we measured shows how we did it. Every answer comes with its confidence, so your customers can decide how much to trust each code, and that's an easier promise to keep than "always right".

For your customers, it's the same arithmetic as for anyone mapping in-house, with made-up numbers as an example: a broker that quotes 30,000 vehicles a month on five insurers and spends 3 minutes on one pair in 50 is spending 150 hours a month on mapping. That's the time they'd be paying you to save.

Getting started

Run the free comparison first, on your own customers' vehicles. Send about 300 recent RC lookups (vehicle fields only) with the codes your customers use, if you have them. Within 3 working days you get our pick, confidence and alternatives for each insurer, and you can judge it before you offer it to anyone.