Gary Saranrota

And although retail and insurance sound wildly different at the core, it's

If I make this decision, what's going to happen?

And these are the core decisions in retail.

Merchandise planning is the core, kind of weekly promotions, prices, inventory allocations.

And also, you know, what to promote, what not to promote, that important mix.

And then in insurance, it's underwriting, claims processing, fraud detection, those kind of core things.

Yeah, we get their transaction log receipts.

So literally every item and every transaction, bricks and mortar and e-commerce.

So the goriest detail.

I know every single item and every transaction for several years.

And then around that is all what promotions they did, what were the prices they charged.

you know, all of that kind of operational details.

It's like the laws of physics, right?

If you think of an autonomous race car, there's no historical data how to drive a lap around a racetrack.

You're not building a predictive model.

So you have the laws of physics and you say, okay, I can simulate the laws of physics.

And then the historical data just kind of configures the simulation, like the gravity, wind resistance, friction of rubber on the road.

So the same way we use the historical data to calculate...

cannibalization, halo sales, elasticity, and then we plug it into our laws of physics, which is this kind of theory of retail.

And similarly, we have a theory of risk.