Jun 26, 2025 · 30m · a16z
Can AI Solve Healthcare's Urgent Workforce Challenges? w/ Ankit Jain
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this episode of the a16z Raising Health podcast, host Julie Yoo interviews Ankit Jain, CEO of Infinitus Systems, on how specialized AI voice agents and copilots address healthcare workforce shortages by automating complex administrative tasks.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. How this is scored →
The host as informed peer 4.3 Guest teaching 4.5 Guest disagreement 1.5 The host pushing back 1.6
speaking balance: gold is the host, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 24:21 Refusing tech budget classification
Ankit firmly rejects the host's framing that their software is viewed as a technology purchase, clarifying immediately that they have never been part of the IT budget.
Hardest push from the host ▶ 18:20 Challenging defense against Big TechJulie confronts the guest with the existential investor question, pushing him to defend why major players like Google, Amazon, or OpenAI won't commoditize his business.
Biggest teaching moment ▶ 21:40 Exposing systemic human error ratesAnkit educates the host on healthcare operational realities by revealing that calling the same payer twice yields different answers 25% of the time, demonstrating why their AI now pushes back on human representatives.
The host holds their own ▶ 7:32 Identifying inverse product trajectoryJulie demonstrates strong industry knowledge by pointing out that starting with fully autonomous agents and adding copilots is the exact reverse of how most AI startups scale.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | The host as informed peer | Guest teaching | Guest disagreement | The host pushing back | Why |
|---|---|---|---|---|---|---|
| How Infinitus Uses LLMs to Solve Workforce Shortages | 3 | 3 | 1 | 1 | Julie opens by grounding the interview in Infinitus's early 2019 founding pre-LLM hype and asks about initial human reactions to voice bots. Ankit provides early background and banters lightly about parents trying to build voice agents, describing their initial tests using fake patient names with United Healthcare. | |
| Scaling Calls and Implementing Safety Guardrails | 4 | 4 | 1 | 1 | Julie asks about use case boundaries and risk mitigation, comparing proactive notifications to pizza tracking. Ankit details their scale of 5 million calls and explains using small language models alongside LLMs for technical and human guardrails in high-stakes healthcare scenarios. | |
| Balancing Autonomous AI Agents and Copilots | 5 | 5 | 2 | 2 | Julie observes that Infinitus went from autonomous agents to copilots, reversing the typical market trajectory, and asks about bot-to-bot interactions. Ankit explains how their FastTrack copilot handles hold times and notes that bot-to-bot calls are an unfortunate necessity due to slow API adoption by enterprises. | |
| Navigating Model Evolution and Modular Architecture | 4 | 4 | 1 | 1 | Julie invites technical details on model selection and avoiding vendor lock-in. Ankit outlines their architecture, explaining how discrete action spaces developed in 2019 allow them to easily swap underlying models while maintaining compliance. | |
| Recruiting Mission-Driven Tech Talent in Silicon Valley | 5 | 5 | 2 | 3 | Julie poses the classic investor question asking why tech giants like Google or OpenAI cannot easily replicate Infinitus. Ankit explains that success in healthcare relies on deep workflow integration, SOP customization, and proprietary data rather than core model commoditization. | |
| Selling to Operations and Discovering System Inconsistencies | 5 | 6 | 2 | 2 | Julie asks about go-to-market strategies, pilot structures, and navigating Chief AI Officers. Ankit reveals that early pilots uncovered a 25% discrepancy when calling the same payer twice, prompting them to build AI knowledge graphs that actively challenge human errors on calls. | |
| Tapping Labor Budgets and Solving January's 'Blizzard' | 4 | 5 | 2 | 2 | Julie asks if Infinitus is purchased through technology budgets or labor budgets. Ankit clarifies that they have never been part of IT budgets, explaining how operations leaders use AI agents to handle seasonal surges like January's benefit verification blizzard. | |
| The Future AI Landscape: Super Apps vs. Specialized Agents | 4 | 4 | 1 | 1 | Julie asks about market landscape consolidation between super apps and specialized point solutions. Ankit discusses the competing efforts between EHRs, platforms, and AI startups to own the integration layer for personalized patient care. |
Statements from this episode (23)
Jain tested early voice AI by calling UnitedHealthcare as 'Bruce Willis'
“Before we incorporated Infinitus we built a demo that could call United Healthcare and say, hi, I'm calling to check benefits for patient Bruce Willis.”
Infinitus AI has completed over 5M calls and 100M audio hours
“So just to put this in scale, we've done over five million phone calls, over a hundred million hours of audio, of conversations between machines and humans.”
Jain: People demand perfection from AI instead of human-level baselines
“People want perfection out of technology rather than comparing it to the human counterpart.”
Jain: Infinitus AI performs significantly better than human healthcare counterparts
“So like any other technology system, we're not perfect. We're significantly better than human counterparts.”
Jain: Administrative delays stop 50% of chronic patients from starting therapy
“Almost 50% of patients who get on therapy, or who are supposed to get on therapy, a script is written, never get on that therapy because of the complexities and delays in those systems, which leads to a lot of avoidable medical cost that ends up in the ecosyst…”
Infinitus FastTrack AI navigates IVRs and waits on hold for employees
“So we have a co-pilot called FastTrack that goes through the IVR systems, waits on hold, and then drops in our customers, employees when the other person is ready to talk.”
Jain used his personal cell for hundreds of thousands of AI calls
“For the first three years of the company, that callback number was my cell phone. Right? And we were doing hundreds of thousands of calls. So I was getting 50 to 60 calls a day for various reasons.”
Jain: Major health payers built dedicated call centers for Infinitus AI
“That led us to have a lot of partnerships with the largest payers in the country, largest PBMs in the countries, where some of them have dedicated call centers just for EVA.”
Infinitus retrieves 30% to 40% of healthcare call data digitally
“We've probably gotten rid of a few single digit percentage of phone calls entirely, but in a pretty good percentage of calls 30 to 40% of data we can now get digitally instead of needing on the phone.”
Jain: AI systems should communicate via APIs, not spoken English
“My thesis has always been, why should you have two machines talk to each other in English? They should talk to each other in bits and bytes.”
Jain: Healthcare enterprises find updating IVRs easier than building external APIs
“The reality is for some large enterprises, it's easier to update their IVR systems than it is to create an API, go through info security, and expose an API to an external party because they've never done that.”
Infinitus fine-tunes LLMs on hundreds of millions of labeled healthcare utterances
“The underlying infrastructure we have is focused on being able to rip and replace any model, the best one that's out there, and sometimes use multiple models from different vendors after fine tuning them with our hundreds of millions of utterances that are lab…”
Jain: Healthcare AI buyers care about patient outcomes, not LLM selection
“Our customers don't care what models we use. They want to make sure they can deliver a better, faster, and more proactive experience to their patient populations.”
Jain: Silicon Valley talent is shifting from ads to healthcare tech
“And it turns out Silicon Valley Is getting an injection of people that have spent their careers working on ads and games that as they go through their careers go, I want to do something that matters to me.”
Jain: Recruiting AI talent is getting easier as foundation tooling improves
“Getting the best talent is important. I think it's getting easier over time. I think it's because every major Company building in AI, the foundation models, is making it easier to fine tune, making it easier to bring your data, which becomes a moat, to be able…”
Jain: Core AI will commoditize; value lies in workflow and proprietary data
“Building the core technology is going to be commoditized over time, but actually delivering the value by being part of the workflow is where there's a lot of challenges and being able to use proprietary data to tune it.”
Jain: Calling healthcare payers twice yields different answers 25% of time
“So you call the same payer twice for the same patient, and you will get different answers 25% of the time.”
Jain: Payers correct errors 70-80% of time when AI pushes back
“And so we built a knowledge graph based on all the calls that we do, to know what right looks like, so that now our AI agents, if they hear something that seems wrong, will push back and say, can you check that again? And we find 70 to 80% of time when we push…”
Jain: Infinitus supports nearly 45% of Fortune 50 companies
“I think we support something like 44 or 45% of the Fortune 50.”
Jain: Infinitus sells to healthcare labor budgets, not IT budgets
“We've never been part of the technology budget.”
Jain: Healthcare RCM groups spend half their workday waiting on hold
“Many RCM groups spend more than half their day waiting on hold.”