🚀 Executive Series - From Broken Property Data to Trusted Underwriting Decision: How to Fix Insurance Decisions at Source

Hey it’s Fabio here,

In today’s Executive Series, we’re joined by Anthony Peake, CEO & Co-Founder of Intelligent AI.

Across the market, property underwriting still runs into the same bottleneck: insurers are making big decisions on incomplete, inconsistent, and often inaccessible property data. Intelligent AI was developed through Lloyd’s Lab and has built its proposition around high-resolution property risk intelligence, rebuild cost analytics, and what it describes as a broader “digital twin of risk” for insurers, brokers, and property portfolios.

So we focused on a practical question: what does it actually take to move from fragmented property data to a more complete, trusted, and AI-ready view of risk?

Anthony’s argument was simple: insurers do not have an AI problem first. They have a property data problem. Fix the data foundation, and better underwriting, claims, and portfolio decisions become much easier to achieve.

The big challenge I saw was that insurers are underwriting with only a fraction of their data. Even five years later, many are still in the same position.

Our thesis has always been delivering a digital twin of risk. In property, there’s a model called COPE — Construction, Occupancy, Protection, and Exposure. There are about 100 pieces of data insurers rely on. Where is it? What is it? How tall is it? What’s its footprint? What’s the wall made of? What’s the roof made of? What type of protections are in place?

Even today, most insurers are still making underwriting decisions on incomplete property data.

So we developed the platform to automate the unlocking of those data points and give insurers a much fuller view of the asset they are underwriting.

I called the company Intelligent AI because we try not to use AI in the wrong way.

We consult the client because we’re dealing with a regulated industry. That means being able to audit and check every piece of data. But we do use AI to unlock the data in the first place and to build models around that data.

So for us, AI is not about adding noise. It is about extracting usable information from unstructured sources, cleaning it, normalising it, and making it useful inside regulated underwriting and claims workflows.

Instead of relying on partial submission data, you are building out a fuller model of the property - not just the obvious physical data, but the wider exposure around it as well. Environmental data is not just earth, wind, and fire. It can also be: this is a perfectly safe office building, or it is next to a higher-risk neighbouring exposure.

That is the difference between reading a file and actually understanding the risk.

Publicly, Intelligent AI describes its platform as a 360° property-risk view that combines geocoding, satellite imagery, flood maps, construction details, fire-risk data, crime analysis, document reading, and digital twins of property risk.

To me, there is huge value in the data. And now, with the advent of AI to unlock that data, measure its accuracy, and use cloud for fast processing, we’ve reached a sort of perfect storm of all of these technologies.

We can give insurers the tools to pre-fill the portfolio so they can make better risk-management decisions, underwriting decisions, and claims decisions. 

The next stage is not just giving people better data, but giving them solid ground-truth data that people can start to train agentic engines with.

The key point is trust. Lots of clients say they have good data, but can you trust it? We always give clients an algorithmic confidence, and we give clients the background of where we get every data source from.

We’ve launched our Risk API in the UK. We’ve announced the partnership with Guidewire. We’re also taking what we’ve built in the UK into the US market.

The data challenges we’ve solved in the London market are just as relevant in the US.

The BCIS (Building Cost Information Service) partnership gives us trusted ground-truth cost data and lets us pre-fill what used to be manual. The next stage is giving insurers a stronger data foundation that is accurate enough to support the next generation of underwriting AI.

The takeaway from Anthony is clear: insurers have a property data problem before they have an AI problem. Fix the trust layer first. Then AI starts becoming useful inside real underwriting and claims workflows.

Fabio Caravita — say hi on Linkedin
Founder, AI Insurer Brief

AI Insurer Brief is 100% FREE for the AI Insurance Community.
If this issue was useful, help us grow by sharing it with 2 people who would genuinely read it. Click below:

Reach insurance execs and decision makers weekly.
72% Exec / C-level. We hit a 50%+ open rate every week.
Limited number of partner spots open each quarter.
→ Reply “Partner” for details and we’ll get in touch.