Use case · High volume
Pricing that runs itself, at aggregator volume
Motor, home, pet, travel, gadget and embedded cover. Swallow serves deterministic prices over APIs, aggregators, embedded forms and AI in milliseconds, with no one in the loop
Thousands of quotes, each one exact
Every request runs your published model, applies its rules and comes back with a price or a decline. Each quote is kept with its inputs, result and model version, so you can see what was offered to whom
Live quotes · todaylive
Quotes1,284,590
Declined by rules96,236
People in the loop0
TimeChannelProductPrice
09:41:16.8AggregatorHome463.33
09:41:16.3Voice AIPetDeclined
09:41:15.9ChatGPTPet108.62
09:41:15.5ChatGPTPet122.48
09:41:15.1Embedded formGadget11.72
09:41:14.6APIGadget13.42
09:41:14.2Embedded formHome487.12
09:41:13.8ChatGPTHomeDeclined
What it takes
Small premiums, big volumes, no room for error
Commoditised products win on price, speed and reliability. The rating engine has to be fast and exact, and changing it has to be safe
Deterministic
One answer, every time
A price depends only on its inputs and the live model version. Ask again next week and you get the same number, which is what aggregators, regulators and finance all expect.
Fast
Milliseconds, even at peak
Comparison site traffic arrives in bursts. Swallow keeps returning prices in milliseconds while millions of quotes a month go through it.
Automatic
Every quote settled on the spot
Declines, refers, excesses and endorsements are rules in the model, so each quote is decided the moment it's asked for. Nobody needs to look at it.
Thin margins
Rate changes that are cheap and safe
A travel or gadget policy can't carry a pricing team. Changes have to be quick to make, tested before they go live and easy to undo.
How it works
One model, served everywhere
Build the model in Swallow from your rating guide or spreadsheets, or with our team. The same published model then serves every channel, and every change is tested first
Serve
Aggregators, your website, partner apps and AI assistants all call the same published model, so the price on a comparison site matches the one on your own site
Try · UK Pet InsuranceRun quote
{
"species": "cat",
"breed": "persian",
"pet_age": 3,
"region": "greater_london",
"vet_fee_limit": 4000,
"excess": 100,
"neutered": true
}Calculated result255.34Valid
Change safely
Every rate change runs against the model's tests before it's approved, including the quotes that should decline. You see which prices moved and by how much, and the previous version is a click away
Tests · 5 of 5 as expectedRun all tests
TestResultValidMovement
persian_london255.34True0%
labrador_north269.76True0%
senior_cat213.56False0%
frenchie_refer621.47True0%
mismatch102.14False0%
Built for scale
Live on price comparison websites today
50M+quotes served, with millions more every month
3M+quotes a month for our biggest customer
Liveon price comparison websites, at peak traffic
0people needed to settle a quote
Two ways to price
High volume or commercial?
One engine prices both. What changes is where the inputs come from and who makes the final call
Commoditised, high volume
Personal lines and embedded cover- Inputs
- Structured, from forms and APIs
- Who decides
- The model's rules, with no one in the loop
- The price
- Final, and the same every time
- What's kept
- Every quote, with its inputs and model version
Commercial and speciality
Fleet, property, liability and speciality- Inputs
- Broker emails, schedules and loss runs
- Who decides
- An underwriter or broker, working in an AI client
- The price
- A technical price for the underwriter to work from
- What's kept
- Every quote and what-if, each with its full trace