There is an invisible industry that sits between an American patient and the treatment their doctor prescribes.
Someone has to confirm what the patient’s insurance will actually cover. Someone has to get the insurer’s permission before treatment can start. Someone has to enroll the patient in financial assistance and follow up on refills. So every year, millions of patients leave the doctor’s office with a prescription they never start. Not because medicine failed them - but because the healthcare system did.
Almost all of this still works the way it did in 1995 - someone at a pharmacy or hospital grinding through hold queues, fax machines, portals, and paperwork, one patient at a time. That labor is baked into what every American pays for healthcare, and it’s why treatment that should start in days sometimes takes weeks.
Infinitus was founded to help close this gap - and make healthcare access, adherence, and affordability easier than ever.
Watch the full podcast episode by scrolling to the end of the article.
Ankit Jain first came up with the idea for Infinitus while at Google, in 2019. At the time, Google was demoing Duplex, an AI that could make calls - like to book dinner reservations at a restaurant. He mentioned this to his wife, who works in healthcare. Her reaction: “You have this incredible technology that can make phone calls on someone’s behalf - and it sounds like it can be used for a lot more than just booking restaurants.”
Ankit was inspired to reach out to a longtime friend and collaborator, Shyam Rajagopalan - a systems architect who had built high-throughput infrastructure at Snap and Google and designed CPUs at MIPS and NVIDIA earlier in his career. The two of them decided to run a test, to see if they could apply this new conversational AI technology to a much more complicated space: healthcare.
Ankit and Shyam called UnitedHealthcare and recorded a robotic voice saying, “Hi, this is Eva calling on behalf of Bruce Willis, and I’d like to collect benefits.” The first agent hung up immediately - but the second one answered every question “Eva” asked, before eventually saying, “Sorry, Bruce Willis isn’t a patient in our registry.”
That was when it clicked - this could work. Infinitus was born.
Infinitus revolves around one of the “least forgiving” interfaces in software - a live conversation, where there is no time to think, and no undo button.
If the AI takes two seconds too long, the person on the other end says “hello?” and hangs up. If it gets a coverage detail wrong, a patient waits longer for treatment. And these conversations can run past an hour, with hundreds of back-and-forth exchanges, which is more than most models can reliably handle by themselves.
To resolve this, Infinitus is built with the following:
A patent-pending Agent Response Control (ARC), which dynamically determines how tightly an AI agent's responses should be controlled during a conversation
A map of how every insurer actually operates, built from millions of real interactions - including which number to call, how each phone tree runs, and what each plan covers
A real-time pipeline - where turning speech into text, deciding what to do, and speaking back all happen seamlessly in milliseconds
What Infinitus is building next is an agent that stays on a patient’s case for months, across channels – and this is opening up exciting new technical challenges for the team to tackle.
Much of the foundation already exists. Every interaction getting traced, scored, and fed back into the system, so the stack gets smarter with use. A workflow layer coordinating agents across a patient’s case, deciding what happens next and when a person needs to step in. Voice, chat, and SMS all running on the same core, so a conversation can move between channels without losing anything. Safety rules computed per customer and per regulation, checked in real time inside every conversation. A model-agnostic AI layer that picks the right voice, speech, and language model for the moment, instead of locking the whole system to one vendor. Memory that carries a patient’s context across calls, programs, and time. A quality system that grades and improves itself through constant simulation, evaluation, and optimization. And soon, AI agents and workflows adapting each patient’s journey in real time.
Infinitus is now supporting over half of the Fortune 50 healthcare companies. Their agents have handled more than 350 million healthcare interactions and 10 million calls - saving countless hours for patients and providers.
This year, Infinitus’s AI beat human benchmarks by more than 18% on many tasks; over 90% of patients have rated the interactions positively; and admin costs are getting cut by up to 30%. Fast Company named Infinitus one of the world’s most innovative companies of 2025, and Andreessen Horowitz led the company’s $51.5M Series C, bringing the team’s total funding to $102.9M.
Infinitus is a flat organization that “doesn’t do titles.” There are no managers, senior managers, directors, or VPs.
If you’re an individual contributor, your title is a “member of X team.” If you lead a few people, you’re the descriptor of what you lead, with the word lead. (For example, Ankit himself carries the title of “Company Lead” rather than CEO. His co-founder’s title is “Technology Lead.”)
The team follows five main principles that they call “CODES”:
Customer-obsessed
Optimistic realists
Disagree and commit
Engage employees
See-say-do – own the problem, don’t just spot it
In the hiring process, Ankit personally interviews everyone. He looks for people who have agency and don’t wait to be told what to do - they just step up and do it. People who have urgency and a sense of fearlessness, and who enjoy being thrown curve balls and figuring it out.
Ankit says the best part of his job is “watching people step into work they didn’t think they were ready for, and rise to the occasion anyway.”
The team is based in San Francisco - where the engineering, product, and design teams work in the office three days per week. The GTM and operations teams are spread across the country, serving customers where they are.
Infinitus is hiring across the board. These are the team’s P0 roles right now:
Forward Deployed Engineer (SF / NYC) - Embed directly with Infinitus’s most strategic customers to drive transformational AI adoption.
AI full stack engineer (SF) - Design, build, and scale AI-powered, user-facing products that make Eva’s healthcare interactions more intelligent, intuitive, and reliable
Watch the full podcast episode here:
This newsletter is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. Furthermore, this content is not investment advice, nor is it intended for use by any investors or prospective investors in any a16z funds. This newsletter may link to other websites or contain other information obtained from third-party sources - a16z has not independently verified nor makes any representations about the current or enduring accuracy of such information. If this content includes third-party advertisements, a16z has not reviewed such advertisements and does not endorse any advertising content or related companies contained therein. Any investments or portfolio companies mentioned, referred to, or described are not representative of all investments in vehicles managed by a16z; visit https://a16z.com/investment-list/ for a full list of investments. Other important information can be found at a16z.com/disclosures. You’re receiving this newsletter since you opted in earlier; if you would like to opt out of future newsletters you may unsubscribe immediately.
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.