Mark Varley

speaker

154 appearances 2 recordings 1 series first heard Jul 2026 last heard 11 Aug

Mark Varley’s voice in public audio — every appearance, attributed to the second.

Trend

recordings per month · last 12 months

Recordings per month over the last 12 months — 2 in all, peaking in Aug 2026 with 1.

Appearances

newest first · ▶ plays the moment

Making Risk Flow | The Future of Insurance · Why Better Data Helps Underwriters Say 'Yes' More Often | Mark Varley & Simon de la Hoyde · 14 Jul 2026 podcast

I guess that's what we provide is we want to be agnostic of the way in which people want to consume the data, but really doubling down on the data itself and making sure it is as good as it can be.

Yeah, that's correct.

And again, beyond even AI is what we're also seeing is we've seen a real evolution in the last kind of 12, 18 months of a lot of customers moving and developing their own data teams and actually consolidating around a couple of different tools.

So the big ones being Snowflake or actually what we see more prevalent in the insurance space is Databricks.

So we see a lot of our customers developing Databricks integrations and using that as a modeling tool.

And again, that's quite data hungry.

So being able to provide to our customers, again, that cleanse curated data in ways in which they need it.

So whether they want to be trying to train your pricing algorithm or whether they want to be using it for underwriting decisions for us, it's just the channel.

I think probably not dissimilar to yourselves at Cytora.

We see ourselves as a headless platform where you can build your own work experiences on your own user experiences on top.

Yeah, I think it's about setting the appropriate guardrails.

So one of the things when we started looking at AI, we've been using AI for development for kind of three, four years now.

But as an interface, it's something we've really been exploring about the last six months.

So we have two customers at the moment who are kind of trialing an MCP interface that we've built.

And when I started looking at this with Tom, our CTO, kind of I was the view of it.

Well, you know, the Gen AI can call an API.

So why do we need an extra layer of interface?

What's the value that gets added?

And actually what we've learned through the proof of concept that we're doing and we're evolving into an offering is actually the MCP server is a really useful channel in that it allows us as the custodians of that data to be able to establish guardrails and tell and inform and prompt the AI to not go off.

So when we first started, we plugged it in and we thought, this is fantastic, but it was going off.

Showing 61–80 of 154 · page 4 of 8