Hubi Zieliński (@hubizet) header
Hubi Zieliński (@hubizet) profile picture

Founder CoraAI 💜 I help doctors focus on patients, not paperwork with #AI 🚀 | 2x Exit

Kraków
Joined June 2014
Photo by Hubi Zieliński (@hubizet)Photo by Hubi Zieliński (@hubizet)Photo by Hubi Zieliński (@hubizet)Photo by Hubi Zieliński (@hubizet)Photo by Hubi Zieliński (@hubizet)Photo by Hubi Zieliński (@hubizet)Photo by Hubi Zieliński (@hubizet)Photo by Hubi Zieliński (@hubizet)Photo by Hubi Zieliński (@hubizet)Photo by Hubi Zieliński (@hubizet)
A 2026 JAMA study associated AI scribe adoption with 16 fewer minutes of documentation per 8 scheduled patient-care hours. Sounds modest? Small savings across hundreds of shifts matter. But you still need to measure the result in your own clinic.
1
4
One of my favorite Cora AI customer results isn't a flashy model benchmark. At Symferia, a physiotherapy clinic, the team reports getting around 50 hours per month back from documentation. That's time clinicians can actually feel.
1
7
Shifting training away from recall makes sense as AI grows, but clinicians still need to spot AI errors in real time. Teaching critical appraisal and workflow judgment is more urgent than ever. teapaw.org/JAMA_current/status/21…
💬 Viewpoint by Jesse Burk-Rafel, MD, MRes, and Robert Wachter, MD: In the era of #AI in #medicine, #physician training may need to emphasize bedside clinical skills, management reasoning, and critical appraisal rather than encyclopedic recall alone. The proposed educational impact is a longitudinal approach to training physicians to work effectively with and without #AI. ja.ma/4zEuhcr
1
Burnout rates by specialty rarely tell the whole story. The day-to-day workflow, documentation burden and EHR friction can matter more than the field itself. Fixing those is just as important as specialty choice. teapaw.org/imghelpinghand/status/…
WHICH MEDICAL SPECIALTIES HAVE THE MOST STRESSED DOCTORS? According to the 2025 AMA physician burnout data: 1. Emergency Medicine - 49.8% 2. Urological Surgery - 49.5% 3. Hematology/Oncology - 49.3% 4. Obstetrics & Gynecology - 45.7% 5. Radiology - 45.2% 6. Family Medicine - 45.0% 7. General Surgery - 43.8% 8. Cardiology - 43.5% 9. Gastroenterology - 43.5% The least burned-out specialties included: - Infectious Diseases - 23.3% - Ophthalmology - 25.8% - Pathology - 28.3% Almost 42% of physicians reported at least one symptom of burnout. Would you still choose your specialty if you knew its burnout rate? #MedEd #Doctors #MedicalSpecialties #PhysicianBurnout
7
A transcript isn't a clinical note. A transcript captures speech. A clinical note structures relevant history, findings, assessment and plan. And a clinician still needs to catch missing or invented details before signing off.
2
81% of physicians in an AMA 2026 survey reported using AI in professional practice. But adoption isn't the same as trust. The real opportunity is building tools clinicians trust enough to use in actual workflows.
3
Matches what I hear from clinicians. Nobody wants AI making the call. They want it to take the typing off their plate, so the person making the call has more attention left for the patient. teapaw.org/HealthcareAIGuy/status…
Americans are generally open to AI in healthcare but want humans making final calls
4
I've seen how a polished AI scribe demo can hide the real work that follows. If I were choosing one for a clinic, I'd ask these six questions first. (1/6)
1
5
4. Review privacy and governance. Who has access? Where is data processed? How long is it kept? What does the contract actually say?
1
5. Check EHR fit. 6. Keep clinician sign-off. The goal is a usable draft inside the real workflow, reviewed by the clinician. Not another admin task. (6/6)
My vote: implement it, with one condition. Residents still write the assessment and plan themselves. Let AI draft the history and admin parts, and keep the clinical reasoning with the trainee. That's where the learning happens. teapaw.org/NEJM/status/2118705042…
As a medical director, would you recommend implementing an AI scribe for residents’ clinical notes or recommend that it not be implemented? Read the case vignette and options and vote: nej.md/47WS6QA
2
The worst AI automation in a clinic is one that adds another tab. Copy the transcript. Paste into a chatbot. Fix the result. Paste it into the EHR. You might have automated the easiest part and kept the hardest.
1
8
If you're a doctor building your own AI scribe with ChatGPT, I respect the initiative. But a working prompt is a prototype. Patient-data controls, error checks, permissions and EHR integration are what make it a usable clinical workflow.
1
38
Clinic owners: before buying an AI scribe, do the math. 10 clinicians × 15 minutes saved/day = 2.5 staff-hours/day. That's a hypothetical, not a product claim. Count review time, corrections and EHR steps before calling it ROI.
1
10
AI scribes often default to safe templates, which can flatten clinical nuance. The real value is in reducing admin work, without erasing the reasoning that makes each note useful for care. teapaw.org/IDdocAdi/status/211997…
Unpopular opinion: using AI scribes for notes makes the good note writers worse and bad note writers passable to make it all a homogenous mush of words. A note should intend to communicate thought process. Homogenization makes it all sound the same. We can do better.
1
13
We've turned clinical documentation into a compliance exercise. Doctors spend more time satisfying the system than actually helping patients. AI should change that, not just help us generate even more paperwork.
I asked a few physician friends: Is there more scrutiny on you from your hospital if you're good or bad at caring for patients? Or if you document your notes correctly? The answers were universal: "There are way more eyes on whether I'm billing right."
14
Every AI scribe demo should answer one question: Will reviewing and correcting this draft take less time than writing the note myself? If the answer is no, faster generation isn't the win you think it is.
22