The Story
A company that decided paperwork was a clinical problem
Here is a fact about American cancer care that sounds like a bureaucratic footnote but is actually a matter of days: before a patient can start treatment, someone has to convince an insurer to pay for it.
That someone is usually a human being at an oncology practice, sitting in front of a payer portal, re-typing clinical facts into forms, waiting on hold, and re-submitting when a field is wrong. The process is called prior authorization, and it is one of those things that everyone in healthcare agrees is broken and almost no one has managed to fix. RISA Labs, a Palo Alto company founded in 2024, has decided to treat this administrative delay as what it functionally is - a clinical problem - and to attack it with software.
The pitch is deceptively simple. In oncology, the argument goes, the bottleneck between a patient and their treatment is frequently not medicine. It is coordination. Every payer has its own rules. Every rule has its own form. Every form feeds a queue, and queues in cancer care are measured in days that patients do not necessarily have. RISA's co-founder and CEO Kshitij Jaggi puts it plainly: "In oncology, delays are costly for everyone involved, but especially for patients."
So RISA built a thing to work the queue. The thing is called BOSS - Business Operating System as a Service - and the company is unusually insistent about what it is not. It is not, in Jaggi's words, "a chatbot or a thin copilot." It is not a single model you talk to. It is closer to an operating layer that sits underneath the clinic's existing systems and does the tedious part.
How the machine actually works
The mechanism is worth understanding because it is where RISA's bet lives. BOSS takes a big, messy workflow - say, getting a chemotherapy regimen authorized under Medicare Part B - and decomposes it into micro-tasks. Each micro-task is small enough to be handled by an agent: sometimes a large language model reading a clinical note, sometimes a digital twin simulating a patient's coverage and pathway, sometimes a reinforcement learner that has figured out how a particular payer's portal behaves. The agents pass work between one another, and the system connects directly into electronic medical records and payer portals rather than asking a human to be the integration layer.
This is a meaningfully different architecture from the two obvious alternatives. One alternative is more people - staffing your way out of the backlog, which does not scale and which every practice already knows is expensive. The other is robotic process automation, the brittle screen-scraping bots that break the moment a payer redesigns a web form. RISA's answer is to make the automation intelligent enough to adapt, and to keep the humans for judgment rather than data entry. The company's own framing is that it decomposes complex workflows into micro-tasks "then delegates them to a network of intelligent agents - LLMs, digital twins, and reinforcement learners."
If that sounds academic, it is because it started that way. Before BOSS was a product, it was a paper. Jaggi and co-founder Kumar Shivang co-authored work on a "Digital Twin Ecosystem in Oncology Clinical Operations," which laid the conceptual groundwork for the commercial platform. Most research papers stay papers. This one turned into a company that now says it processes thousands of cancer cases a day.
The numbers, with the usual caveats
RISA reports the kind of metrics that make investors lean forward. At one large oncology practice, the company says administrative staff time dropped substantially, denials fell by as much as 40 percent, and first-pass authorization approval rates climbed toward 98 percent. Authorizations that once took days are now filed within 24 hours, and some sites are approaching turnaround measured in hours. Jaggi's summary version: "We're saving two to four days on every treatment we touch."
The honest caveat is that these are company-reported figures from deployment sites, not an independent study, and "as much as 40 percent" is doing some quiet work in that sentence. But the direction is the point, and in oncology the direction is unusually legible. A day saved is not an abstraction. It is a day a patient is in treatment instead of waiting on a form.
Why two strategic investors both wanted in
In January 2026, RISA closed an $11.1 million Series A co-led by Cencora Ventures and Optum Ventures, with participation from Oncology Ventures, Z21 Ventures, and John Simon through Ventureforgood. That followed a $3.5 million seed round in 2025 led by Flipkart co-founder Binny Bansal, bringing total funding to roughly $14.6 million.
The identity of the Series A leads is the interesting detail. Cencora sits near the center of pharmaceutical distribution and specialty pharmacy. Optum sits near the center of, well, a great deal of American healthcare. Strategic investors of that weight do not co-lead a round for a copilot demo. They co-lead when they think the thing might become infrastructure - a layer other people build on and route work through. That is a bigger and riskier bet than a productivity tool, and it is the bet RISA is asking the market to make.
The stated plan is to scale BOSS nationwide, with an early ambition of reaching 100 cancer centers. RISA is currently live at eight oncology practices, with 15 more under contract, and its published roster of partners and customers reads like a tour of community oncology: New York Cancer & Blood Specialists, Tennessee Oncology, Astera Cancer Care, OneOncology, Mary Bird Perkins Cancer Center, New York Oncology Hematology. Community oncology is where most American cancer care actually happens, and it is also where administrative staffing is thinnest - which is either a coincidence or the entire market.
The people, and the pattern
Jaggi and Shivang are IIT Kanpur alumni and repeat founders who have been friends for more than a decade. Before RISA they built and scaled Urban Health together, which means the current company is not a first rodeo but a second act with a shared operating memory. There is a recognizable founder pattern here: pick an industry where the incumbents are large and the workflows are ugly, go deep on the specific vertical rather than selling a horizontal tool, and treat compliance as a feature rather than a chore. RISA advertises HIPAA, SOC 2, ISO 27001, and HITECH compliance the way other startups advertise features, because in healthcare that is exactly what they are.
What makes RISA worth watching is not that it has automated prior authorization - plenty of companies claim some version of that. It is the architectural insistence that oncology operations are too complex for a single model and too dynamic for a dumb bot, and that the right answer is an orchestration layer of many specialized agents. As Jaggi says, "Oncology is one of the most complex operational environments in healthcare." RISA's wager is that complexity is not a reason to avoid the problem. It is the moat.
Whether that wager pays off depends on things no press release can tell you: whether the agents stay reliable as payers change their rules, whether eight live sites become eighty without the quality quietly degrading, and whether "operating system for oncology" turns out to be a real category or a good slide. But the company has picked a problem where being right matters in a way that is easy to describe and hard to fake. In oncology, the product either gets patients treated faster or it does not. That is a clarifying kind of business to be in.