Thought Leaders in Artificial Intelligence: Daisy Intelligence CEO Gary Saarenvirta (Part 3)

Gary Saarenvirta: We have a theory for retail. We have partial differential equations (PDEs) that govern insurance and risk industries.

We simulate with our PDEs. You simulate what is going to happen in the future if I took these actions. In the same way, our PDEs are the decisions we provide. What products do I promote? What prices do I charge? How much inventory to allocate?

We can simulate what will happen if you did this combination of action. We find the optimal combination of products, prices, and inventory that maximizes retail sales over the course of a year.

We do the same in insurance. We simulate our PDEs to find the optimal prices for policies. What’s the optimal level of auto-adjudication and human adjudication and optimal level of auto-deny to maximize company loss ratio.

It’s the engineering approach, which is first-order closed loop control theory. It has a set of inputs, a system dynamics, an output, and a feedback loop.

We take historical data about the system. In retail, we take transaction details. We have every single receipt, every product bought in every store. We use that to calculate some of the fundamental features. We fit those into our set of PDEs and then we simulate to find the optimal decision.

We solve our equations using biomimicry methods. Those are embarrassingly parallelizable. If you have 100,000 products and you have to choose 2,000 to promote, 100,000 choosing 2,000 is 10 to the power of 10,000. It’s more than the number of molecules in the universe.

The way to search as much of that space as possible is to use some method that is highly parallelizable. That’s where the biomimicry method allows us to massively scale. We do that with GPU computing to search as much of that space as possible to find better decisions than human beings can do today.

Sramana Mitra: When you look around, what are white spaces to you? If you were starting a company today, what kind of things look like interesting problems worth solving?

Gary Saarenvirta: The climate is huge for taking our approach to look at the system and find the key fundamental impacts and features and then build from mathematical theory. I think the climate is a huge human problem. Healthcare is another huge area.

Healthcare is the largest budget line item in the United States. Changing the way healthcare works from being reactive to proactive and improving people’s health so we can minimize the cost. 50% of healthcare spending is spent in the last six months of someone’s life. That is very reactive.

We need to become more efficient or our economics are going to fail. Big human problems is where I would start a business today.

Sramana Mitra: Thank you for your time.

This segment is part 3 in the series : Thought Leaders in Artificial Intelligence: Daisy Intelligence CEO Gary Saarenvirta
1 2 3