Co-founder and CEO of Qumis - the Chicago AI platform teaching software to interpret the dense, technical fine print that runs the trillion-dollar insurance business. He used to do it by hand. Then he stopped.
There is a version of the insurance industry that still lives in a three-ring binder. A young lawyer gets handed a policy thick as a phone book, a fistful of highlighters, a pad of Post-it notes, and one instruction: mark it up. Find the gaps. Find the exclusions. Find the sentence on page 84 that quietly undoes the promise on page 3. Dan Schuleman did exactly that in his first week of practice. He remembers the instruction word for word.
Today he is the co-founder and CEO of Qumis, an AI platform that does the marking up at machine speed - and, more importantly, explains its reasoning the way a coverage attorney would. The pitch is not "robots replace lawyers." The pitch is closer to: the gold standard for reading a policy has always been a skilled coverage attorney, and you simply cannot staff one on every account. Qumis is how you clone the judgment without cloning the person.
That distinction matters to Schuleman, because he was the person. Before he was a founder he spent years inside the machinery of insurance disputes - representing carriers, arguing over what a clause really meant, learning that in a state-regulated business the difference between covered and not-covered can hinge on a single comma. The frustration was not the difficulty. It was the waste. The same dense documents, parsed by hand, over and over, by an aging workforce he describes as still working "from the typewriter era."
Qumis ingests the messy reality of insurance - policy towers, quotes, binders, endorsements - and returns source-linked, reasoned answers. Not a summary. A position you could defend.
Analyzes layered programs and surfaces coverage gaps across complex, stacked policies that humans lose track of.
Lines up quotes, binders, and endorsements across markets so brokers see what actually changed - and what it costs.
Backs coverage positions with expert-quality reasoning chains and citations, trained on thousands of real-world analyses.
The architecture is the part Schuleman gets animated about: proprietary document processing feeding a multi-stage legal reasoning engine. Every answer arrives with its receipts - source-linked citations and a visible chain of logic - because in insurance, an answer you can't trace is an answer you can't use.
He is careful about the word "replace." Adjusters, he points out, do work "akin to practicing law because they interpret insurance policies." Qumis is not built to fire them. It is built to hand them back the hours they currently spend with a highlighter, so the human can do the judgment the machine still can't. He compares the moment to accounting's arrival of the spreadsheet: the tedious math got automated, and the profession leveled up.
Schuleman didn't arrive at AI from engineering. He arrived from the inside of the problem.
Two decades building and scaling platforms across financial services, logistics, and insurance. Former co-founder and CTO of Newtrul. At Goldman Sachs he was SVP and Head of Application Development for U.S. Deposits, helping launch the multibillion-dollar Marcus platform.
Schuleman on the pairing: Shiv "can build without needing traditional product development bureaucracy - extremely valuable at our early stage."
It is the classic high-functioning insurtech split: the lawyer who has felt the pain in his bones, paired with the engineer who has shipped at bank scale. Coverage knowledge meets Marcus-grade infrastructure. One knows exactly what "right" looks like; the other knows how to build it so it doesn't break in production.
He names his rivals without flinching: "spreadsheets, PDFs, and people's brains. And phone calls." It's the most honest competitor slide in insurtech.
Before he ever pitched a VC, he sat as a judicial extern in federal court - watching how arguments get won and lost up close.
His design taste runs to Apple's intuitive philosophy and Airbnb's Brian Chesky. A coverage lawyer who cares about the product feeling effortless.
The team splits its energy 60/40 - sixty on customer growth, forty on building a technical moat. He'll tell you the exact ratio.
Schuleman's ambition is not a feature. It is a re-write of a verb. He wants to change what it means to "read" a policy - to move an entire industry off paper rituals and onto transparent, reasoned, instantly searchable interpretation. Capture the institutional knowledge that currently walks out the door when a veteran adjuster retires, and put it on tap for every broker, underwriter, and claims professional who needs it at 4:55pm on a Friday.
The bet underneath it all is quietly radical: that the most defensible thing in AI for insurance is not the model, but the judgment baked into it - the thousands of real coverage calls that teach Qumis to reason like the attorney who used to do it by hand. He was that attorney. Now he is building the version of himself that scales.