Dedi Gilad

The through-line

Gilad's constant argument is that telehealth as built has hit a ceiling because it cannot examine anyone. Video and audio have been available for twenty years, and yet "utilization wise people don't use it enough you know five to ten percent of the population are using that after a lot of investment into that space" [1]. His explanation is psychological rather than technical: "Healthcare in general and Primary Care specifically is a lot about psychology and trust," and talking to "a physician that you probably don't know and you will probably never see again without any ability for them to diagnose you to examine you reduces the the level of trust" [1]. The fix he has pursued for a decade is to move the physical interaction itself into the house rather than to accept a thinner version of it remotely: "the concept of title is really to take the physical interaction and shift it or replicate it to the home" [1][3].

What has shifted is scope. The origin was narrow and personal, a four-year-old daughter with repeated ear and throat infections and a father making the same trip to the pediatrician every three weeks [1][4][5]. The first commercial wedge was equally narrow, urgent care at night and on weekends [1]. The trajectory since has been outward and downward in acuity: scheduled primary care, chronic monitoring, hospital at home, and now ambitions toward passive, always-on sensing [1].

On why telehealth plateaued and what changes the curve

Gilad treats the COVID bump as a distortion rather than a proof point. Uptake rose "just because population couldn't go out but post covet you see a back to reality uh going down in terms of utilization" [1]. His counter-metric is engagement, not availability: with an exam capability in the home, "we are seeing five to eight times utilization" [1]. He is careful not to present this as automatic. Adoption is "a behavioral change slowly it takes time it's complex we invest a lot into education into engagement" [1]. The product claim underneath is that one kit covers the full physical exam: imaging of throat, skin, ears and eyes, auscultation of heart, lungs and stomach, and basic vital signs, all captured by a family member and read by a clinician who becomes, in his framing, dependent on the household for its senses, "we are becoming the remote hands and ears and eyes of this clinician" [1][3].

On primary care as the common denominator

Gilad frames TytoCare's strategic choice as a deliberate refusal of the usual startup path. "Usually in healthcare is that you take a very small niche of area very small amount of patient and you're trying to solve a very a restricted problem," and instead the company "took a big bet" on primary care because it is "the common denominator" that fits any household, student, hiker, family with kids, or senior living [1]. He concedes the cost of that choice: it is a wide problem, and "you're not really treating the most cost the most expensive areas" [1]. The payoff is sequencing. Once engagement and adoption exist, additional verticals can be layered on the same installed base, so the home device is infrastructure and asthma, hypertension and pregnancy management are floors built on it [1][5].

His critique of the alternative is blunt and aimed at buyers as much as builders. Companies that started from single conditions, "let's do only diabetic and only hypertension and only this and that," have left health organizations "hammered with tons many many solutions many many platforms very very hard to integrate all of them and it's chaotic" [1]. He also positions the company deliberately away from direct-to-consumer, selling B2B2C through payers, providers and payviders, on the logic that any organization taking insurance risk while delivering care is a natural partner [1][5]. That partner model is what he has discussed alongside health system leadership, including how Highmark Health deploys home-based devices [2].

On moving from episodic care to always-on

He describes a clear progression of care modalities. Episodic urgent care came first, then protocol-driven management for periods of life that need structure, asthma, hypertension, pregnancy, so that the system "can actually not wait for you to be sick not wait for you to be a for this visit to be episodic" [1][5]. The next step he wants is continuous and unobtrusive: monitoring specific populations, "toddlers or young babies or very specific chronic patient over a long period of time passively," picking up "more subtle uh changes whether it's overnight or during the day in a passive way without interfering to the to the lives" and feeding prediction rather than reaction [1]. He identified an acquisition in that space as the vehicle for it [1]. The concrete example of protocol care working is asthma, where the system alerts a clinician to deterioration in a child's condition and can keep a family out of the doctor's office "or even worse without going to see the ER" [1]. He has also spoken about the state of remote patient monitoring and how the virtual visit itself is evolving [3].

On AI, and why the data matters more than the label

Gilad is openly impatient with the term. "I don't really like the term AI people are using it so much without a real credential," though he considers the machine learning genuinely distinctive because the inputs are simple and abundant: images, videos and sound recordings gathered from hundreds of thousands of patients, anonymised and analysed in the background [1]. The first regulatory milestone he cites is FDA clearance for lung sounds analysis, automatically detecting wheeze and crackle [1]. He claims the resulting dataset is not purchasable elsewhere: "you cannot go out to Mount Sinai or to Kaiser Permanente just buy it nobody holds that kind of information" [1].

His argument about what the AI is for differs from the standard one. Most healthcare AI, as he characterises it, sifts large volumes of data to do tedious work quickly and catch what a clinician might miss. TytoCare's approach is to have best-in-class pulmonologists and cardiologists label the data, then push that judgment down to generalists: a pediatrician "might have learned about you know fine crackles in med school but in reality they don't see that phenomenon too often," so tagging by specialists means "giving them the tools of the specialist to the front line of primary care" [1]. He sets an explicit boundary on ambition: "I'm not saying that we will replace clinician we're not even aiming for that we're trying to be very realistic," the goal being insights, alerts and decision support in the near and mid term [1].

Takeaways

• Telehealth's ceiling is trust, not technology: a clinician who cannot examine you and whom you will never see again produces low utilisation, roughly five to ten percent of the population despite heavy investment [1].
• Adding a home physical exam moved utilisation five to eight times higher, but only alongside sustained investment in education and engagement, because the change is behavioural and slow [1].
• Betting on primary care as "the common denominator" for every household is harder and less lucrative up front than a point solution, but it creates the installed base on which asthma, hypertension and pregnancy programmes can be layered [1][5].
• Point solutions leave health organisations with too many platforms to integrate; Gilad calls the result chaotic and sells instead to payers, providers and payviders that carry insurance risk [1][5].
• The care roadmap runs episodic urgent care, then protocol-driven chronic management, then passive always-on monitoring of toddlers, babies and chronic patients for subtle overnight and daytime changes [1].
• FDA clearance for automated lung sound analysis of wheeze and crackle is the first proof point for the AI layer [1].
• Specialist-labelled data is the differentiator: tagging by pulmonologists and cardiologists puts specialist-level interpretation in the hands of front-line PCPs, with no aim of replacing clinicians [1].