Fixing the ASC Cancellation Problem: DataLily AI's Madison Startup Origin Story
- Jul 20
- 3 min read
Updated: Jul 30
Two co founders had been grinding through a retail analytics platform when clarity finally hit. "We spent one night, we did an all-nighter, we had it built; and then we literally asked ourselves, 'Why?'” Sumanth Karnati, DataLily AI Co Founder, says. "We stopped and we were like, this is not what we want to be building."
They pivoted. Healthcare had pulled at them twice before; first at their first hackathon through Open AI at New York Tech Week, and a competition win for improving rural healthcare access for GE Healthcare.
The second pull towards healthcare came later. Sumanth comes from a farming background and wanted to solve something real. Elly Kruse, DataLily AI Co Founder, had worked in her father's healthcare revenue cycle management brokerage firm and had a personal run-in with the system's absurdity: a prior authorization issue left her with a bill for thousands of dollars, a full year after her care.
When they landed in Madison, healthcare was everywhere they looked. They started talking to people: providers, administrators, surgery centers. Within a month of focusing on ambulatory surgery centers, outpatient facilities that are 60-70% more affordable than hospitals and growing fast, they had a design partner.
That's where DataLily AI was born.
The Problem They're Solving
Ambulatory surgery centers (ASCs) are a better deal for almost everyone - safer, cheaper, and more than half are owned by their physicians. But growing one means flying blind: operators guess where to build, guess which surgeons are recruitable, and negotiate against insurers who know what every center in the market gets paid. An independent ASC can collect a fraction of what a facility across town gets for the identical procedure - and have no way to prove it.
Rose, DataLily's ASC market-intelligence platform, gives operators the leverage: billions of records distilled into which markets need a center, which surgeons are worth a conversation, and what insurers actually pay. Its companion platform, Lily, handles the daily grind - flagging surgical risk before it becomes a last-minute cancellation and automating the prior-auth and claims work that buries staff.
Their StartingBlock Chapter
Elly and Sumanth found StartingBlock in the Fall of 2025 through a connection with Eric Leventhal, who works at the space and is a health tech founder. They joined Level Set, StartingBlock's early-stage accelerator, while still shaping their problem statement. The programming helped. The community helped more.
"Before StartingBlock, we were learning off of ChatGPT or watching YC videos. But you couldn't ask it a question and get a real answer," Sumanth says. Elly follows up; "Here, if you have a problem with something, you just connect to the right person. It's like the six degrees of Kevin Bacon game, there's always someone.”
In Fall 2025, they submitted to the Elevator Pitch Olympics at the Wisconsin Early Stage Symposium, pitching among 19 Wisconsin companies. They won. They also took home the People's Choice Award. Among the judges panel and in the crowd that day, they met the investors who would go on to fund their pre-seed round.
Scott Mosley, StartingBlock CEO, handed Sumanth two books: Smart Brevity and Venture Deals. It was the kind of thing you can't get from a chatbot. "Having that human connection, having these connections, we even got some of our customers just through being part of Madworks," Sumanth says.
In Spring 2026, they completed the Madworks Accelerator. Weekly mentor sessions with healthcare operators, health system veterans, and contract specialists gave them frameworks they didn't know they were missing. The one that stuck most? The reverse RFP. "We didn't even know what an RFP was," Elly says. "And then our mentors told us, you combat vendors by telling customers what they should look for. You're doing them a service. That was mind-blowing.”
Elly also credits a less glamorous memory: cold-calling from every corner of the building, doing squats to stay focused. Sumanth remembers napping on a couch after nights of one or two hours of sleep.
They now are working with 15 surgery centers nationally.
What's Next
The five-year vision is 600 ASCs running on DataLily AI end-to-end, making independent surgery centers viable, so surgeons keep their autonomy and patients get the care they need. "The day you think everything is going well is the day when it's not," Sumanth says, "and the day when you think it's not going well is the day when things are actually happening. You just have to believe in what you're doing and keep going."
