Company Deep Dive: Counsel Health

Subscribe | Advertise | AI Startup Database | Feature AI Jobs

TOGETHER WITH COUNSEL HEALTH

Industry Deep Dive: Counsel Health

Perspectives from the people building the future of health AI…

We sat down with Dr. Rishi Khakhkhar, CMO at Counsel Health, a virtual medical practice specializing in messaging-based care. Counsel offers patients unlimited access to medical advice from expert physicians and is redefining what AI-enabled, asynchronous care looks like at scale. A practicing emergency physician and digital health operator, Dr. Khakhkhar shared how his team is rethinking access, continuity, and clinical workflow from the ground up.

Let’s start at the top. What is Counsel Health and how did you get involved?
Counsel is a virtual care company built around asynchronous care and AI-native workflows. We’re not just layering AI onto old processes—we’re building everything from the ground up with these tools in mind. The goal is to explore what AI-enabled care delivery actually feels like day-to-day, not just in theory.

I joined after practicing emergency medicine at Mount Sinai, where I led virtual and home-based care programs. I love being in the clinical trenches, but I’d been thinking about digital health since the early 2010s. When I met Muthu, our CEO, it was clear he had a bold, but physician-centered vision: to use AI to augment clinical decision-making, not replace it—and start with real doctors, delivering real care, from day one.

What’s been the biggest surprise about asynchronous care?
Honestly? How human it still feels.

I used to think you needed video to build connection—that something important was lost without face-to-face interaction. But with async messaging, we’ve actually unlocked something new: longitudinal relationships. We follow up with over 90% of our patients, even for simple issues. Our engineers built a feature to schedule check-ins, and it’s been one of the most powerful tools we have. Patients feel seen. They don’t fall through the cracks.

We’ve also added things like guided video uploads and safety triggers—if I suspect something more acute, I can call a patient directly. Async isn’t a downgrade. It’s a different mode of care, and with the right scaffolding, it can actually feel more continuous and personal.

Where does AI fit into that model today?
AI is built into nearly every layer of what we do. From intake to documentation to decision support, we use frontier LLMs to lighten the clinician’s load. It’s like having a resident who takes the first pass: asking history questions, summarizing the chart, and surfacing key info. It gets me to the decision point faster, and lets me focus on what matters most.

We’re not building our own models, but we are deeply integrated. We use AI to surface relevant literature and plan to expand patient-facing AI features soon. But we’re careful: unsupervised AI can heighten anxiety if left unchecked. We see our role as the translator—meeting patients where they are, then guiding them forward with clinical judgment and empathy.

How do you think about AI replacing vs. supporting clinicians?
We think of AI as a multiplier, not a replacement. There’s so much ambient demand for care—questions patients don’t even ask because they assume they’ll be a burden. Our job is to move from a world of scarcity to one of abundance—and AI is how we get there.

But ultimately, there are still decisions only a physician can make: prescribing, referring, delivering a diagnosis with compassion and clarity. That moment of trust between doctor and patient—that’s still the heartbeat of care. And we think AI makes that moment more accessible, not less.

Counsel describes itself as part software company, part research lab, part medical practice. How does that play out day to day?
It’s one of the biggest shifts coming from a traditional health system. At Counsel, our doctors and engineers work hand-in-hand. We built our own EMR to match the unique demands of asynchronous care. There wasn’t an off-the-shelf solution that worked for our needs, so we created an internal system tailored to the way we deliver care.

The other piece is data. Messaging-based interactions generate real clinical conversations at scale—far more nuanced and longitudinal than static forms or video visits. That gives us a unique opportunity to both study async care outcomes and improve our AI systems in the real world. We believe sharing those learnings—technical and clinical—can move the whole industry forward.

How is AI actually used under the hood?
We’re not training our own foundation models, but we’re deeply integrated with the best frontier models out there. AI touches almost every process internally—from patient intake to clinical decision support to chart summarization. 

The goal has always been to build “Iron Man suits” for doctors: tools that bring them the right context, from the medical record to current guidelines, so they can make faster, smarter decisions. It’s about efficiency, accuracy, and scale. We’re also exploring how AI can surface local referral pathways—matching patients to high-value, in-network care based on their geography and insurance.

Do you plan to automate more of the care experience over time?
Absolutely—but with care. Right now, we’re still very much in a "physician-in-the-loop" phase. Doctors see and edit every response. But over time, as the models improve and our confidence grows, we expect some interactions to be more automated, especially for low-acuity use cases.

Think of it like Waymo was a few years ago: there’s still a human in the seat today, but the long-term vision is autonomy. We see a future where doctors are only looped in when necessary,allowing us to stretch our clinical team further without compromising quality.

How are you scaling the physician network to keep up with demand?
Today, our entire clinical team is in-house. We’ve explored third-party staffing solutions, but asynchronous care is still so bespoke and so different from traditional telehealth that it requires specialized onboarding and tooling. Docs at Counsel deliver care in a way that’s new to them, and it takes time to ramp up. Having our own medical group has helped us maintain quality and accountability as we grow.

What’s the longer-term clinical vision? Will you expand beyond primary care?
Definitely. We started with general medical advice and triage—the kinds of questions people used to take to Google or nurse lines. Now we’re expanding into more longitudinal care: managing conditions like hypertension or high cholesterol, and working toward a concierge-like experience where your doctor, wearable data, and even your trainer or dietician are all connected.

Our aim is to be a personal doctor in your pocket, backed by the reach and intelligence of an entire clinical team.

How are you working with employers, health plans, or health systems today?
Our main go-to-market motion is B2B: employers and payers. We help them expand access, improve member experience, and avoid unnecessary urgent care or ER visits. One of our core pitches is: put a doctor as close to the front of the healthcare journey as possible. That early, ongoing access reduces downstream costs and improves outcomes.

We’re also partnering with health systems on transitions of care. Imagine a patient who gets 24/7 support in the hospital and then nothing the next day – that’s the reality for the vast majority of patients discharged from a hospital today. Counsel can bridge that gap—with daily check-ins and easy access to a doctor for the first 30 days post-discharge. That alone can meaningfully impact readmissions and follow-up compliance.

Healthcare AI Guy Summary

What stood out, what’s tricky, and why it matters…

Counsel Health is building a messaging-first medical group that uses AI to make care faster and more responsive, without removing doctors from the loop. Instead of video visits or form-filling, patients message in, get thoughtful replies, and can follow up over time. It’s like texting your doctor, but with structure behind it.

They’ve built their own EMR, plugged into large language models for things like intake and documentation, and are figuring out where AI helps without making things weird or unsafe. Right now, doctors still review every message. Long term, the idea is to let AI handle more of the routine work so clinicians can focus on higher-touch decisions.

The early focus has been triage, urgent care, and basic primary care. But they’re already moving into post-discharge care and chronic condition management. GTM looks like most healthcare startups: employers, payers, and health systems. They also raised an $11M seed round from tier 1 investors, led by Andreessen Horowitz, so they are well on their way.

What stood out

• Messaging is the main event: Most virtual care still leans on video. Counsel flipped it. Messaging is the core product, not an afterthought. Patients don’t just get one-off responses. They get follow-ups, check-ins, and actual continuity of care. That’s rare.

• AI is in the background, not the brand: Counsel isn’t trying to wow anyone with AI for the sake of it. They use existing models to handle repetitive stuff like chart review and message drafting. It’s embedded in the workflow, not added on top.

• They built their own EMR! Rather than fight with off-the-shelf systems that weren’t built for async care, they made their own. That gives them more flexibility to tweak and improve how care gets delivered.

• They’re not stopping at triage: Counsel started with quick wins like “do I need to go to urgent care,” but they’re already expanding into longer-term relationships. Think managing blood pressure, reviewing lab results, or coordinating follow-up care after a hospital stay.

• Could be the missing piece to the lifestyle medicine trend: Companies like Function Health are exploding and offering comprehensive lab tests and personal insights on folks’ health, but are missing the physician network to help people act on those results. Solutions like Counsel Health might be the missing piece here.

What’s tricky

• Scaling the clinician network: Async care isn’t just shorter visits. It’s a different workflow entirely. Knowing when to escalate or follow up based on a message thread takes training, and that’s hard to outsource.

• Getting patients to use it like chat, not like a portal: Most people still treat contacting a doctor as a serious event. Part of the challenge is helping users treat this more like messaging a knowledgeable friend, not waiting until something feels urgent.

• Keeping automation in check: For now, every message gets a human review. But as they automate more, they’ll need strong systems to make sure nothing slips through. No hallucinated diagnoses. No missed red flags.

Final thoughts

Counsel isn’t trying to reinvent healthcare from scratch. They’re building on what already works, swapping video for messaging, and layering in AI in practical ways. By focusing on infrastructure, clinical continuity, and thoughtful use of AI, it’s shaping what a more scalable, human-centered future in healthcare could look like. We’re excited to watch their journey unfold!