Show Notes
In 2017, while most founders were still debating whether chatbots had a future, Punit Singh Soni was studying speech models with the patience of someone who had already seen what came next. He was not a healthcare guy. He had run games at Google, built mobile apps, and sat in the social team. But he understood one thing the rest of the industry was about to learn the hard way: AI was about to become the new UI, and the biggest unlock would happen wherever sophisticated users were drowning in repeatable, unstructured workflows they hated doing.
That pointed straight at medicine. Doctors had become data clerks. Patients were getting 13 minutes of face time, half of it spent watching their physician type. So Punit founded Suki with a single mission: bring presence back to healthcare. Today, Suki is the ambient clinical intelligence layer running quietly inside Zoom, Optum, Athena, Meditech, and a growing list of healthcare giants, valued at roughly half a billion dollars and built on the contrarian belief that the best product in a regulated, bureaucratic industry is not a feature. It is giving someone their time back.
Frameworks from This Episode
These frameworks have been added to the AI for Founders Frameworks Library. Filter by Healthcare or Punit Singh Soni to find them.
The Four Arcs of Ambient Clinical Intelligence
Punit's mental model for what an AI layer in healthcare actually does, broken into four sequential capabilities.
- •Clinical documentation: capturing what happened in the encounter.
- •Assisted revenue cycle: extracting financial information so the doctor and system get paid.
- •Clinical reasoning: providing contextual information back to the doctor based on patient history.
- •Clinical operations: running agents on the encounter output to automate downstream tasks.
- •All four arcs together create a complete ambient intelligence layer. Most vendors only cover one.
The Android Analogy for Platform Strategy
How Suki structured a dual go-to-market without splitting focus or diluting the core product.
- •The Suki app is the reference implementation. Suki sells it directly to health systems.
- •The Suki platform is the underlying layer, licensed to Zoom, Optum, Athena, and others to power their own clinical AI products.
- •Selling the reference product teaches the company how to build the platform.
- •The platform creates ecosystem footprint without requiring Suki to own every customer relationship.
- •Most founders choose one model or the other. Punit chose both and sequenced them intentionally.
Where AI Will Have the Biggest Impact
Punit's filter for picking a market in the AI era. Look for three conditions at once.
- •A sophisticated user who knows exactly when the AI is wrong and can course-correct.
- •Lots of unstructured data that no existing system captures cleanly.
- •Repeatable workflows the user finds boring or burdensome, where automation has no emotional cost.
- •Healthcare scored highest on all three. That is why Punit started there.
- •Run this filter on your own market before writing a single line of code.
Eating Glass and Love as a Strategy
Aggression and warmth are not opposites. They are the two halves of a sustainable founder operating system.
- •Building requires constant willingness to confront your own inadequacy. That is the eating glass part.
- •Surviving that requires self-empathy, which only comes from extending care outward first.
- •Punit's philosophical framework: dharma is the responsibility to act, karma is the inevitability of outcomes.
- •Kodawari, the Japanese concept of pursuing excellence in a craft for its own sake, rounds out the operating system.
- •Founders who have only aggression or only warmth burn out or burn others. You need both.
The Future Doctor
The doctor's role does not disappear. It shifts. Understanding the shift is how founders build for what comes next.
- •The world has 7 to 8 billion people and not enough physicians to serve them all.
- •AI expands access and personalization simultaneously. It does not replace physicians; it multiplies their reach.
- •Tomorrow's clinician is a student of both medicine and AI.
- •The role shifts from gatekeeper of knowledge to guide who helps patients navigate a sea of AI tools.
- •Founders who build for the gatekeeper model will be obsolete. Build for the guide model.
Founder Experiment
Pick the single most repeatable, unstructured workflow your team complains about. For most founders, it is meeting notes, customer call summaries, or post-call CRM updates. Then build the smallest possible ambient version using the Anthropic API.
Spin up a Claude project that ingests a transcript and outputs four artifacts in parallel: a clean summary of what happened, the action items extracted as structured data, the contextual information someone would need before the next interaction, and the downstream tasks an agent could automatically execute. That is Punit's four arcs applied to your business. Run it on ten real transcripts before you write a single line of UI code. The point is not the tool. The point is forcing yourself to articulate which workflows in your operation are sophisticated-user, unstructured-data, repeatable-burden problems. That is where your AI moat lives.
Glossary
- Ambient Clinical Intelligence (ACI)
- An AI system that passively listens, captures, and processes clinical encounters without requiring active input, then automates downstream documentation, coding, and operational tasks.
- EHR (Electronic Health Record)
- The digital system hospitals use to store patient information, codes, and clinical notes. Often described as a collection of rules and regulations rather than a tool designed for human interaction.
- ICD-10 codes
- The standardized diagnostic codes attached to clinical encounters for insurance and claim processing.
- Revenue cycle
- The financial workflow from patient encounter to payment, including coding, billing, and claims processing.
- Two-pronged sales motion
- Suki's structure of selling the reference app to health systems while licensing the underlying platform to other healthcare technology companies.
- Kodawari
- The Japanese concept of pursuing excellence in a craft for its own sake, without requiring external recognition.
