Hospitals run on data, yet nurses and doctors still spend precious minutes typing notes and chasing stale vitals.
Ashwin Chandramouli, who leads portfolio growth strategy and business development at Baxter International, joins the podcast to explain where AI in healthcare is genuinely helping, where it is not, and why adoption decides everything.
Can AI fix the paperwork crisis in healthcare? Ashwin Chandramouli’s view
Ashwin sees AI as a workforce and workflow productivity tool rather than a replacement for clinicians. It can query data the way a person would, surface what matters and remove the first steps of gathering and structuring information.
He is clear about the limit: a technology only creates value if clinicians adopt it. That means fitting into existing workflows, since adoption drives utilisation, value and then renewals.
- Labour is the biggest pressure. Labour is the top line item on a hospital P&L, and nursing shortages and burnout make it worse.
- Ambient documentation is working. Tools that turn a conversation into structured notes are already common in outpatient care, with a human reviewing the result.
- Real-time data beats stale data. Device data from beds and monitors can supplement vitals that go stale after rounds.
- Adoption is the linchpin. A shiny tool that does not fit the workflow will not be used.
How does a healthcare technology leader think about business and clinical needs?
Ashwin started in investment banking, moved into consumer packaged goods running a P&L, then into healthcare technology at Abbott Diagnostics and Hillrom before Baxter. His father is a radiologist, so healthcare was familiar territory.
His view is that business and clinical thinking go hand in hand. You cannot drop a new technology into a hospital without respecting existing clinical and operational workflows.
Who has to say yes when selling technology to a hospital?
Ashwin describes hospital sales as compartmentalised, with different stakeholders prioritising the same needs differently. Positioning is more art than science.
- Clinical stakeholders: care how the product affects patient care and workflow.
- Operational and procurement teams: ask whether it is standard technology and how easily it plugs in.
- Financial leaders and buying committees: control the purse strings.
Where is technology moving fastest in healthcare?
Ashwin says the fastest movers are tools that resemble what clinicians already use at home, such as voice assistants. A voice-enabled wearable badge his team is launching is one example of this approach.
He notes the commonly cited figure that technology takes about 17 years to become standard of care. Medical devices move slowly because lives are at risk, but software that is not a medical device can move faster.
What can video and AI do in a hospital room?
Cameras with algorithms can understand contextual cues, such as a patient at risk of falling who is trying to get out of bed. An automatic alert could bring someone in sooner, instead of waiting for a button press.
What is connected care and why does real-time device data matter?
Baxter is now mainly a medical device company, with infusion pumps and beds from its Hillrom acquisition. Ashwin describes a data integration layer that pulls information from pumps, monitors and beds and delivers it to clinicians.
In his example, bed sensors can track heart rate and respiratory rate continuously. Charted vitals go stale once a nurse leaves the room, so combining this feed with EMR data gives a more real-time picture.
- Think of it as a real-time operating layer on top of the EMR.
- The EMR, in his words, behaves more like a finance system than a clinical workflow tool.
- Tying a product directly to patient outcomes is hard, because many other factors are involved.
How is AI already used for documentation and revenue cycle management?
Ashwin points to ambient tools such as Microsoft’s Nuance DAX, Abridge and Suki, widely used in outpatient care. A doctor records a visit on a phone, and the tool turns it into structured notes for the EMR, with the provider checking accuracy.
He says this differs from older dictation, which is verbatim. Saved time lets doctors review insights and think about treatment instead.
- Capture the patient conversation with an ambient tool.
- Structure it into a note and push it into the record.
- The clinician reviews and corrects it.
- Cleaner notes improve clinical documentation integrity.
- That speeds payer authorisation and payment in revenue cycle management.
What are the biggest unsolved headwinds for US hospitals?
The first is labour. Nurse shortages and burnout create a cycle in which fewer nurses mean heavier workloads and more attrition.
The second is payment pressure. He says new legislation is set to cut reimbursement for many hospital systems, while hospitals already run at high occupancy, so they cannot simply add volume. The focus therefore falls on cost productivity, where AI is expected to help.
How do large device companies organise around connected care?
Traditionally each product line has its own P&L. Connected care cuts across all of them, like the mortar between bricks, so Ashwin says organisations are still figuring out how to manage it.
His advice is to assume positive intent, build trust and collaborate across business units. He sees connectivity as a strategic way to build a durable software revenue base alongside capital equipment sales.
Further reading and sources
- Baxter International
- Baxter International (Wikipedia)
- Revenue cycle management (Wikipedia)
- Electronic health record (Wikipedia)
Frequently asked questions
How is AI used in healthcare documentation today?
Ambient tools record a patient visit, turn it into structured notes and send it to the electronic medical record. A clinician still reviews the note for accuracy, which saves time compared with typing or verbatim dictation.
Will AI replace nurses and doctors?
Ashwin Chandramouli does not expect that soon. He sees AI as a productivity tool to help overstretched staff cover workload gaps, as long as the information it surfaces is trustworthy.
Why do healthcare technology products fail to get adopted?
According to Ashwin, many tools are introduced as shiny objects that do not fit the existing workflow. Without adoption there is no utilisation or value for customers, and without value there are no renewals.
Conclusion
Ashwin’s message for healthtech builders is to fit the workflow first. The best AI in a hospital saves clinicians time, earns their trust and connects to how the hospital actually makes money.
If you sell into healthcare, map every stakeholder in the buying chain and design for adoption before you design for features.
Want more conversations like this? Join The Builders Club, our community of 35,000+ founders and CXOs, and explore more episodes of the Startup Founders Podcast.
