How 8 Healthcare Services Startups are Using AI to Get Ahead
In “Services: The New Software,” Julien Bek of Sequoia argues that the next wave of AI winners will come from companies that sell outcomes rather than tools. In other words, the value is shifting away from selling software and toward delivering the service itself.
In that sense, healthcare is already there. The industry has always been built around companies that provide care, with third-party software increasingly layered throughout. Many of those tools have been valuable, but they are starting to look interchangeable, whether because incumbents (ahem, Epic) replicate and underprice key features, or because it’s becoming increasingly easy to build them in-house.
That shift puts healthcare services startups in a strong position. They have something pure software plays do not: a full stack. They own the care model, the patient relationship, the clinical data, and the operations. When AI is embedded across that stack, it strengthens the core service and reinforces a competitive advantage.
In a previous piece on tech-enabled healthcare services, we laid out a guide for thinking about where to deploy tech and AI in your startup. This post is the companion to that, with eight real examples of healthcare services companies putting AI to work in ways that already move the needle.
1. Honor: cutting caregiver turnover with smarter scheduling
Home care has an annual caregiver turnover rate of 85%. Honor has made a real dent in it by using AI to figure out what care professionals actually want from their schedules, and then making sure they get it.
The system maps stated preferences (days, hours, client types, location) against revealed preferences derived from actual behavior, since the two often don’t align. It then matches caregivers to clients in a way that works for both sides. Churn for care professionals in this cohort has dropped to the mid-30s.
Shout out to Honor CEO Seth Sternberg for giving me the idea for this article!
2. Ro: faster response times for patients reporting side effects
When patients message about medication side effects, speed matters. Ro built an in-house AI triage tool that identifies side-effect reports, classifies their potential severity, and automatically routes them to the right care team member.
Before the tool, Ro was already responding in under two hours. After deployment, median response time dropped to 33 minutes across all hours of the day. Urgent messages get answered in under 26 minutes. That’s a 70%+ improvement, and the data has been peer-reviewed and presented at Obesity Week 2025.
3. Akido: enabling same-day treatment in street medicine settings
Akido is using its AI-native system to power a national medical network, including its street medicine program serving unhoused people in California. In this program, Community Health Workers use their proprietary AI system, ScopeAI, to guide structured intake, surface targeted follow-up questions, and generate real-time clinical summaries for physician review.
This enables community health workers to expand their patient capacity while handling more clinically complex interactions, all under the supervision of a licensed provider.
In Akido’s street medicine program, patients average 2.6 visits per month, with 63% retained at six months (which is unusually high for this population). Meanwhile, emergency department visits have dropped by 31%, pointing to a model that expands capacity while driving positive health outcomes that help prevent emergencies and save costs.
4. Everlywell: increasing patient follow-through on at-home tests
Everlywell is using Eva, its AI health companion, to turn a fragmented healthcare experience into a more proactive, coordinated one. Eva interacts directly with consumers, answering questions in real time, anticipating needs, and guiding them to next steps based on their context.
For example, if a member receives an A1c kit but hasn’t returned it, Eva initiates a call that feels like a natural conversation rather than a phone tree, introduces itself, verifies the member, and explains why the test matters. Eva resolves nearly 70% of inquiries on first contact (roughly on par with trained human teams, at a fraction of the cost) and drives higher engagement and completion rates, especially among Spanish-preferred members.
5. Aledade: improving care decisions at the point of care
Primary care physicians make dozens of decisions per patient visit. Most EHR systems show them what’s in their own records. Aledade’s EHR Overlay to Aledade Assist™ pulls in labs, pharmacy data, health information exchanges, and claims data, then surfaces AI-enriched insights directly inside a clinician’s existing workflow at the moment they need it.
Since launching in 2024, more than 1,000 primary care organizations have implemented it, with over 85% of eligible practices activating the tool. Clinicians using it to identify diabetic patients who might need a statin saw a 40%+ relative improvement in 14-day prescription fill rates. Medication reconciliation completion rates for older adults rose by more than 10%.
6. Virta: improving patient outcomes through personalized guidance
Virta’s metabolic care program helps patients shift their metabolism to burn fat for energy, but what stands out is how their AI makes that approach actionable in everyday life. Members track what they eat alongside real-time biomarker data from connected devices like glucose and ketone meters, allowing the system to learn how each individual uniquely responds to carbohydrates.
Their AI system, Metabolic Intelligence, connects these inputs to outcomes, helping members understand what’s working and where they’re getting off track. It then translates those insights into practical next steps, including personalized recommendations and alternative foods that better support their goals, all tailored to individual preferences and routines.
Members using the system are about five times more likely to make that metabolic transition. Those who do are 1.67 times more likely to lose at least 5% of their body weight within six months, and on average lose twice as much as those who don’t.
7. Origin: reducing documentation burden for clinicians
Documentation is one of the leading causes of clinician burnout, and AI scribing tools are multiplying fast. Origin built its own web extension that sits on top of its EHR, compiles patient history, assists with note writing, and surfaces payer requirements in real time.
After a documentation refresh in Q1 2026, clinician NPS jumped more than 50 points. As off-the-shelf scribing tools become more standard, companies that have deeply integrated documentation AI into their specific workflows will hold an edge, at least for a while.
8. Cityblock: automating patient outreach at scale
Cityblock has used AI to return more than 17,000 hours to its care team. For example, their Personalized & Automated Voicemail feature lets AI take the wheel when member calls go to voicemail (which happens about 75% of the time). Previously, outreach specialists spent upwards of eight hours a week on a manual, script-bound process. Cityblock’s AI-generated voicemails are tailored to each member based on their care plan, recent interactions with their care team, and open needs.
The result: more than 5,000 hours saved, with specialists reclaiming roughly 25% of their call time, or about 75 minutes per day.
Bringing it all together
Healthcare services companies have historically been penalized by investors for thin margins and high labor costs. But when AI meaningfully reduces the cost to serve, it changes the margin profile. And better margins create room to reinvest in what actually improves care: better clinicians, better tools, and a more consistent patient experience.
That’s what makes this moment interesting. AI can mean faster time to treatment, more personalized care, higher patient engagement, and the ability to reach populations that the system has historically struggled to serve. In many cases, AI is helping close the gap between knowing what to do and actually doing it by turning fragmented workflows into something more proactive and coordinated.
That’s what I see as the real opportunity right now. Healthcare services companies that own the full stack are in a uniquely strong position to compound these gains over time. Lower cost to serve, better outcomes, and stronger patient experiences reinforce one another. That’s what ultimately makes these businesses more durable, both clinically and financially.
The pure software plays are fighting over an increasingly crowded and price-sensitive market. The healthcare services companies that get this right are playing a different game entirely.
