Leveraging AI to Reduce Churn and Increase NRR - with Dan Harmeson, Co-Founder and Co-CEO at QuadSci

**Ray Rike** (0:08)
Welcome to the AI to ROI podcast. On today's episode, I am joined by Dan Harmeson, the co-founder and co-CEO of QuadSci.
We'll be covering three primary topics with Dan today but at a fairly detailed level. First, what are the key factors in successfully deploying AI to manage churn, and expansion in a B2B software company? Second, how is AI being used, and one of the best practices to predict churn using product telemetry data? Third, how is AI being used to increase that holy grail of recurring revenue, software metrics, net revenue retention? So Dan, with that, can you please take a moment to give a brief overview of your journey to becoming a guest here on the AI to ROI podcast?

**Dan Harmeson** (1:00)
Well, thanks for having me, Ray. Pleasure to be here. Sean and I started QuadSci in August of 2023 Sean Murray is my co-founder and we wanted to bring scientific management to the world of B2B software.
The world of B2B software has this incredibly data, powerful data advantage where people are paying to use subscriptions and in the process of that, they're generating all of this telemetry data. So that's like a product analytics, front-end UI, UX, navigation clicks, even some of the agentic front-end interactions, all the way to backend service level observability metrics. So that entire space of telemetry data contains very predictive and descriptive signals that allow you to better understand your customer base. So we wanted to have a flavor of machine learning that deeply understands those patterns and unlocks those for go-to-market. Telemetry data is actually the single biggest data set that B2B software companies have.
It's just not used for go-to-market. So that's what we made Cloud Sci for.

**Ray Rike** (2:04)
Remember when cloud costs first started surprising people?
Finance would open the AWS invoice. No one could explain what drove it. And engineering would spend a week building a spreadsheet to explain it. And it was still wrong.
AI is doing that again, except faster, more dynamic, and spread across more systems and departments. Today, a single enterprise has multiple AI costs running across multiple vendors such as AWS, inference costs on Anthropic and OpenAI, GitHub co-pilot cursor, and a handful of AI agents that no one billing system sees the entire spend picture. So finance asks, what did we spend on AI this month? And the answer takes three days to find out, and it's still probably wrong. And what did we get for it? Most don't even try to answer that one. That is the problem Maverick was built for. Maverick gives finance IT in engineering a single source of truth for AI spend. So you can allocate costs, enforce budgets and connect investments to business outcomes. Learn more at maverick.ai. That's M-A-V-V-R-I-K dot AI. Now on to the show.
Hey Dan, for the non-technical people out there, would you mind just kind of defining what telemetry data is so they understand it?

**Dan Harmeson** (3:20)
Yep, so you go on to a website, you log in to an app, you click that button, you fill that form, that generates something called product analytics events. So UI, UX.
If you're in a coding console or even one of the more modern GenAI tools and you prompt or you configure an API, that fires a service or a function in a software product. And then you go follow that down and that generates something called an observability metric or an observability trace. And so these product analytics events from front end, these deeper service level functions, observability metrics and traces, that's what constitutes these billions and in some cases trillions of events for every B2B software company that they have around how their customers are actually using their products.

**Ray Rike** (4:06)
Yeah. When we were first talking, preparing for today's episode, we talked a little bit about your experience at the SAS Institute and what you did there. Then you were the head of customer success at Elastic. Can you just tell me a little bit about what you did and what you learned there that became the catalyst for applying machine learning to reducing churn and increasing net revenue retention?

**Dan Harmeson** (4:30)
Yeah. QuadSci is just a natural combination of Sean and I's experiences. Sean ran sales teams at Salesforce. He ran services teams at MuleSoft. He ran CS at Elastic.
When we were at Elastic together, I ran strategy and operations for go to market. When I was at MuleSoft, I ran a services organization. When I was at SAS, I ran a front office skunk works of going and finding new use cases for applied machine learning out there in highly regulated industries like banking, insurance, manufacturing, supply chain. What we've seen through a combination of our experiences is go to markets very well-trained in the world of B2B software. There's a lot of rigor, there's a lot of process, but it's really hard to have precision and really knowing what your customers are doing and is it good or bad. And meanwhile, customers are having greater and greater expectations of the service that they get, of the interaction that they get from a B2B company, especially one that they're spending a lot of money on. Everybody's using Amazon on the home front, right? Everybody's using YouTube. Everybody's using these tools and apps that have these hyper-sophisticated user experiences, right? Netflix movie recommendations. Why don't I get that stuff in my B2B products?