Thank you for being one of the 355 subscribers who’ve waited patiently for this moment - the first official edition of Health AI Foresight.
I had a peek through the subscriber list and saw some of the most prominent healthcare and AI leaders among you - this is going to be fun!
Before I dive in, a quick intro: I’m Dr. Youssef Aboufandi - a son of a patient who struggled to navigate healthcare, and the reason I became a physician in the first place. Before training and practicing across Medicine, Surgery, and A&E in the NHS, I spent a year at IQVIA, where I saw how research, data, and technology can transform healthcare globally.
Now, I merge both lenses - clinician and analyst - to track healthcare AI’s evolution: the early signals, the emerging trends, and foresight into what’s next before it unfolds.
Let’s dive in.
In short, the latest signals:
EHRs are becoming AI-native platforms
AI-CDS tools are reaching clinicians faster than expected
OpenAI is moving in on both clinicians and patients
The “AI Doctor” era has begun
Wearables are evolving into a Personal Health OS
⏳ If you’ve got five minutes - read on.
Signal 1: EHRs are becoming AI-native platforms
What’s happening
Epic rolled out its AI roadmap: AI Scribe (via Microsoft Dragon Copilot), Emmie (patient-facing AI), Art (clinician copilot), and Penny (revenue cycle management). Its Cosmos AI now spans 115 billion medical events across 118 million patients, powering predictive models of future health outcomes. 160+ AI projects have have been announced.
Oracle Health unveiled its AI-native EHR which is voice-first, flexible, and built for multimodal model integrations.
Other EHRs are following fast: Athenahealth revealed its AI-native athenaOne EHR, combining AI scribing, revenue-cycle automation, and workflow tools; and eClinicalWorks launched its AI-Powered EHR, embedding generative models directly into charting and note review.
Foresight:
EHRs are no longer static infrastructure - they’re becoming cognitive systems. Startups can’t rely on the old line that “EHRs don’t innovate” anymore. The edge now lies in integrating deeply into workflows and solving narrow, high-value pain points that EHR giants can’t replicate without structural change.
Signal 2: AI-CDS tools are reaching clinicians faster than expected
What’s happening
The AI scribe is evolving from administrative relief into a true decision-support interface.
Doximity acquired Pathway for $63 million, embedding evidence-based CDS directly into DoxGPT.
OpenEvidence released a transparent-reasoning scribe that cites every source.
Wolters Kluwer launched UpToDate Expert AI, built on decades of peer-reviewed content across 7,600 physicians with explainable reasoning.
Foresight:
Ambient scribes + AI-CDS tools are merging into one category - the clinical copilot. Adoption is accelerating fastest in the U.S., while the U.K. and Europe lag behind. With UpToDate’s trusted brand and global hospital footprint, it’s positioned not only to compete with Doximity and OpenEvidence in the U.S., but to dominate internationally. Expect AI-CDS to reshape workflows and empower allied health professionals to take on larger roles in a physician-short world.
Signal 3: OpenAI is moving in on both doctors and patients
What’s happening
OpenAI has quietly built a dedicated healthcare team while pitching GPT-5 to hospitals and marketing patient-facing chat experiences.
New hires include Nate Gross, MD (Doximity co-founder) to lead healthcare go-to-market, and Ashley Alexander (former Instagram co-head of product) as VP of Health Products.
Foresight:
ChatGPT today reaches around 700 million weekly users globally. Even if only a small fraction use it for health-related queries, that still represents tens of millions of patients engaging with it in medical or medical-adjacent ways. That reach alone could become its moat.
With its growing hospital partnerships and expanding model capabilities, OpenAI may eventually connect both sides of care - patients and providers - through a single model infrastructure, positioning itself as a potential gateway to care. It will focus on becoming the middleware where data, reasoning, access, and infrastructure begin to converge.
Signal 4: The “AI Doctor” era has begun
What’s happening
After ChatGPT’s rise, patients realised they could query health issues instantly - and the market responded.
Doctronic launched an AI doctor app offering free consultations and $39 human-doctor escalation, providing symptom assessments, possible diagnoses, and care plans.
There are also more subtle patient-facing copilots like Epic’s Emmie - not an ‘‘AI doctor’’, but a conversational assistant embedded directly within hospital systems. Emmie can explain lab results, answer questions, and guide patients between visits - yet given its integration with real clinical data, it has the potential to evolve into one.
K Health and Akido Labs are blending AI with clinician oversight - their models use AI to suggest likely diagnoses, while human doctors validate and manage care. They’re not fully autonomous or patient-facing, but I placed them in this list as they signal how AI tools are beginning to take on more of the diagnostic workload.
Plenty of unregulated patient-facing AI Doctor apps have already emerged and some are currently being built in stealth.
Foresight:
This isn’t “wellness AI.” It’s the emergence of clinical-grade patient intelligence - tools that lower access barriers and take on triage, self-care, and low-acuity cases. They’re the smarter, contextual successors to Dr Google - but built for action rather than search. I foresee healthcare providers beginning to recognise these tools as powerful patient-facing gateways. Over the next year, expect LLM copilots to be embedded directly into hospital systems - guiding patients from their first query at home all the way to inpatient supervised care pathways.
What’s emerging is a new form of patient empowerment -where AI becomes the first point of contact, extending clinical capacity beyond the walls of hospitals. As regulation matures and trust builds, the idea of an “AI Doctor” will move from a futuristic concept to an accepted part of everyday care.
For systems where readmissions directly drive costs - like the NHS - this shift could be transformative: Patient-facing AI copilots can flag risks early, support adherence, and manage minor conditions before they escalate into avoidable admissions.
Signal 5: Wearables are evolving into a Personal Health OS
What’s happening
Ōura and WHOOP now offer in-app lab testing through Quest Diagnostics - results reviewed by clinicians and displayed directly in the app.
Ōura Health Panels: 50+ biomarkers linked by an AI Advisor to sleep, stress, and recovery.
WHOOP Advanced Labs: 65 biomarkers feeding recovery and training algorithms.
Partnerships snapshot: Quest & Ōura, Quest & WHOOP, Labcorp & Life Extension, Function Health & Equinox. All converge on one loop: continuous data + labs + AI coaching.
Foresight:
Consumers now control their biology from the wrist - a huge step for prevention. The next advantage isn’t more data; it’s connection. Whoever unifies these partnerships into a seamless experience - or owns the full stack - will define the Personal Health OS. Interoperability is what’s missing now and what everyone is waiting for.
As wearables, EHRs, and AI agents converge, raw health data will turn into actionable intelligence serving both patients and care teams.
That’s a wrap for Edition #1 of Health AI Foresight.
My goal with this newsletter is simple: to surface the early signals shaping healthcare’s AI future - and offer foresight into where it’s heading next.
If this helped you see the direction more clearly, share it with one colleague or team trying to make sense of this fast-moving space.
See you in the next edition,
- Dr Aboufandi
📩 Reply if something stood out - I read every message.