Mark Dangelo in Rick’s Room

This episode of Rick’s Room features Mark Dangelo, a longtime technology leader and industry thinker with deep experience across mortgage servicing, financial services, healthcare, and telecommunications.

The conversation centers on a critical issue facing modern organizations: how to effectively access, govern, and use data in the age of AI.

Dangelo explains that organizations are generating massive volumes of data—on the order of 10²³ to 10²⁴ units annually—yet much of it remains underutilized. 

While business intelligence tools promised better insights decades ago, many companies still struggle to extract meaningful value.

AI dramatically increases output speed, but poor-quality or poorly governed data leads to bad decisions at scale.

A data foundry standardizes and defines data once, enables reuse across multiple use cases, and supports regulatory compliance and reporting.

According to Dangelo, AI is like a high-performance jet engine—but if you feed it low-quality fuel, it will fail catastrophically.

Podcast Highlights

Rick Grant:
Hi, and welcome back to Rick’s Room. This is the show where we simulate that end-of-day lounge at a business conference—the place where you finally get to talk to the people you really wanted to see.

Today, my guest is someone I’ve known for more than 30 years. He’s a longtime technology leader, a published author, and a deep thinker on data and systems. He’s worked across multiple industries and has written extensively for the Mortgage Bankers Association and Thomson Reuters.

Mark Dangelo! Mark, thanks for being here.

Mark Dangelo:
Thanks, Rick. It’s a pleasure to be here.

Rick:
You’re one of my favorite people to interview because you always make me think differently. Let’s start here: what have you been working on lately?

Mark:
I’ve been focused on a fundamental question:
How do organizations actually access and use their data, especially in the age of AI?

This goes back before generative AI. Even in the early days of machine learning, the question was always:
How do we use data to create predictability?
How do we generate consistent outcomes?

If we can unlock that data, we can uncover insights organizations didn’t even know they had.

Rick:
And there’s a lot of data out there.

Mark:
A massive amount. Industry estimates suggest we’re generating on the order of 10²³ to 10²⁴ units of data every year.

The real question is: What happens if we can unlock even 10–15% more of it?

That’s what I’ve been working on, how to do that without adding more complexity, more software, or ripping out existing systems.

Rick:
That’s interesting, because companies have been trying to solve this for years with business intelligence tools. And many still struggle.

Now with AI, the risk is even greater. If the data is bad, the output gets worse. So how do we fix that?

Mark:
It starts with governance and access.

Here’s a simple truth: No one understands data better than the people who create it.

So instead of centralizing everything, we need to go back to the source. That’s where the concept of a data mesh comes in.

A data mesh means managing data within the business units where it’s created, leveraging the expertise already in the organization.

Rick:
So you’re not trying to move all the data into one place?

Mark:
Exactly. We don’t want to replicate everything or build another warehouse. We want to leave the data where it is and understand how to access it properly.

Then we connect those domains using a data fabric, which applies rules around access, updates, and integrity.

And finally, we introduce the idea of a data foundry.

Rick:
Tell me about that.

Mark:
The data foundry is where everything comes together.

If we understand our data—where it comes from, how it’s used, and how it’s governed—we can define it once and reuse it across the organization.

For example, in regulatory reporting, instead of submitting slightly different data to multiple agencies, we define it once, submit it once, and ensure consistency.

Rick:
That’s a big shift.

Mark:
It is. And it’s necessary, especially with AI.

In the past, we copied and moved data over and over again. In an AI-driven world, that approach is dangerous.

AI systems—especially agentic AI—are making decisions. They need clean, traceable, reliable data.

Rick:
Otherwise, you’re scaling bad decisions.

Mark:
Exactly.

Think of AI as a high-performance jet engine. If you feed it low-quality fuel, it won’t just underperform. It will fail.

Rick:
Who needs this most right now?

Mark:
Highly regulated industries—financial services, healthcare, any organization dealing with multiple regulatory bodies.

Because if bad data gets into the system today, it doesn’t stay contained. It spreads, and fast.

Rick:
That’s a powerful insight.

Mark, every time we talk, I feel like we’re just getting started. We’ll have to have you back.

Mark:
I’d love that. Thank you.

Rick:
Thanks for joining us. And thanks to everyone watching. We’ll see you next time.