24-07-2025
The AI Will See You Now: Clinical Care Agents At Your Service
Insight app.
Going to the doctor is a tedious experience at best—and a traumatizing one at worst—so unpleasant that many people avoid it altogether, until an emergency or crisis occurs. The reality is all too familiar for most Americans who can expect to wait weeks or months for an appointment, sit in the waiting room for hours before seeing a provider, fill out redundant medical questionnaires, and then have a rushed interaction with a busy nurse or doctor who barely listens to your health concerns.
Meanwhile, the administrative burdens surrounding clinical care have led to provider burnout and workforce shortages within the medical profession. As reported in Harvard Business Review, U.S. physicians spend 34% to 55% of their workday on documentation and EHR reviews—cutting into face time with patients and weakening the patient-provider relationship.
I spoke with the team at Insight Health, a startup that has raised $4.6 million in seed funding from investors including Kindred Ventures, RTP Global, Wedbush Ventures and MKT1, to alleviate the routine aspects of care across the patient journey—screening, intake, documentation, and follow-up—using AI clinical care agents.
Insight Health founding team - Jaimal Soni (Chief Executive Officer), Dr. Pankaj Gore, Dr. Eric ... More Stecker (Chief Medical Officer), Saran Siva, (President)
'I don't think the problem can be overstated,' shared Dr. Eric Stecker, who is a co-founder and the Chief Medical Officer. 'Practicing physicians often feel like we're banging our heads against the wall, trying to fix a broken system.' And for our clients at Insight Health, the screening or intake process is a real pain point, elaborated co-founder and CEO Jaimal Soni. A typical neurosurgery visit could be 30 to 40 minutes, with half of that time spent capturing a patient's history, including family, medical and surgical history. And the clinicians often run out of time.
Since its founding in 2023, Insight Health has conducted over 100,000 autonomous clinical conversations with patients, deploying its agents to support thousands of clinicians mostly at mid-sized private practices and community hospitals but also at large institutions across various specialities, including Neurosurgery, Oncology, Gastroenterology, and Primary Care.
How Does The AI Platform Work ?
I was able to visit Insight Health's office in New York City for a live demo, and chat with their AI agent 'Lumi' during a screening process via phone call. Lumi has a number of modules, or workflows which can be personalized and leveraged by clinicians to engage directly with the patient, and also speaks multiple languages.
First, it integrates with a practice's EHR and referral notes, to extract relevant data about the patient (i.e., the details from the patient's last clinic visit or last annual checkup). Then, it reaches out to the patient over email, text or phone for a screening visit, in order to gather an updated patient history. Following the screening, Lumi generates a structured clinical summary for the provider to review before the in-person visit. Additional workflows support appointment scheduling, in-office visit summaries via an AI scribe and post-visit feedback collection.
The accuracy of the audio-to-written transcription which was generated during the Q&A phone screening process, was noteworthy, but I wondered what would happen when a patient went 'off-script'.
Dr. Stecker shared that 'with generative AI, we can program it to provide clarification, to identify inconsistencies within a patient's medical history and to explain basic medical terms, but not give medical advice.'
Are Patients Willing To Share Health Data With AI?
A key consideration, particularly with older patients, is whether they would be comfortable speaking to an AI about sensitive health information. Dr. Stecker remarked, 'I've been surprised by how well the technology has been received. One patient spent a long time interacting with the AI, sharing detailed information about their pain and medical history — you could tell they really needed someone to talk to.'
Pre-visit summary
Research has shown that people like sharing personal details when the AI is relatable and empathetic.1 A recent study in ScienceDirect showed that 'individuals were just as likely to choose to self-disclose to an AI as to a human researcher'.2 And this study built on previous research that found that self-disclosure to chatbots, text or voice based AI agents who engage in naturalistic simulated conversations,3 evokes similar emotional and psychological satisfaction compared to human-human interactions.4
But Is It Safe?
Insight Health is HIPAA and SOC 2 Type 2 compliant and protects PHI, or personal health information consistent with industry standards, however Soni noted that in the US, the regulatory body around AI for healthcare just doesn't exist yet.
The current administration has unwound certain HHS policy efforts that were put in motion by former President Biden. For example, Executive Order 14110: Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence was signed into law by former President Biden in October 2023, establishing standards for AI use in healthcare and other industries. But, the order was rescinded by President Trump on January 20, 2025.
The One Big Beautiful Bill Act which was signed into law July 4, 2025 did leave out certain 'moratorium' provisions that would have prohibited nearly all state and local regulation of AI. Meaning, if we do see guardrails placed around artificial intelligence in healthcare, it will likely be through state and locally mandated requirements.
So in lieu of an established regulatory infrastructure, Insight Health has built an internal safety framework called 'Safe AI' to quantitatively measure the quality of outputs, identify outliers and escalate issues to clinicians. Essentially, the platform processes every transcript to verify accuracy and look for anomaly detection prior to generating a medical summary.
'We found patients may contradict themselves (i.e., a patient says they are on a diabetes medication but did not mention they are a diabetic) and we highlight this in a note for the clinician,' Soni says.
Will AI Replace Doctors?
As the article title suggests, an obvious question remains – are AI providers going to be replacing doctors, nurses and other healthcare providers?
Dr. Stecker made a distinction between what Insight Health is doing, in supporting or aiding the practice of medicine, and 'making recommendations, giving a diagnosis and prescribing a treatment plan' which is 'getting into the practice of medicine.' But the reality is, this is a slippery slope. 'The lines are blurring as the systems get better,' said Kanyi Maqubela of Kindred Ventures, one of Insight Health's seed investors.
A recent study completed by Cedars-Sinai, Tel Aviv University found that startup K Health's AI physician assistant matched doctors' clinical decisions in two-thirds of patient cases, while offering better care in the remaining third. The research also found that the AI made potentially harmful recommendations 2.8% of the time, versus physicians' 4.6%.5 As might be expected, the AI 'was better at following guidelines but wasn't as good where there was nuance.' Dr. Caroline Goldzweig, Chief Medical Officer at Cedars-Sinai elaborated in an interview with Forbes, 'With complex patients with a lot of comorbidities, that's where you really do need human intervention.'
Maqubela shared that ultimately the patient still wants expert advice coming from physicians and providers, so the challenge is leveraging AI to create 'a wrap-around experience for the human expert, so humans can just do the things that are unique to the patient and everything else can be optimized by AI.'
Perhaps we will soon see AI replacing providers when the case is straightforward and without complication. But for now, platforms like K Health and Insight Health position their respective platforms not as replacements, but as critical tools to support physicians, nurses and other providers with routine tasks, to free them up to provide more higher-level and high touch support to patients.
What's The Future Of AI In Care Delivery ?
Maqubela shared his broad vision for the future of AI in the care delivery space, and furthermore, what gave his venture firm conviction in Insight Health's strategy.
Healthcare 'has historically been a laggard in all software.' But, AI is different. 'We're seeing massive adoption, as the technology can take large volumes of unstructured, multimodal data (including labs and diagnostic images and complex family history), give probabilistic responses, produce summarization, and do robocalling that is personable.' There is incredible market-fit with healthcare in AI that has surprised a lot of folks. 'And we're just in the first inning.'
In phase two, we're evolving beyond scribes and summarizations, and more towards dynamic and robust agents 'that can move from the EHR into the diagnostic, into the LLM, into the voice application, back into the EHR, with the right supervisory systems on top of them.' And Insight Health's founding team includes deep specialists in Cardiology and Neurology (Dr. Eric Stecker and Dr. Pankaj Gore) which Maqubela believes will be critical in training future agents.
