Why we built Medows as a workspace, not an EHR

A junior doctor at one of India's busiest government medical colleges once told me, "I don't need another system to chart in. I need somewhere to think."

That sentence is why Medows exists.

The category most clinical AI products are stuck in

If you list the AI tools doctors hear about in 2026, almost all of them sit in one of two places:

1. Inside the EHR. Documentation copilots like Abridge, Nuance DAX, and Suki transcribe the visit and write the SOAP note. They're sold to hospitals. They make the existing record-keeping system bearable.
2. On the open web. OpenEvidence, Glass Health, UpToDate-style products answer general clinical questions. They have no patient context — the doctor restates the case each time.

Both categories are useful. Neither is what a doctor on the ward is actually doing all day.

What a ward shift actually looks like

If you watch a resident on admission day, you see something neither category serves:

• They're holding twenty to thirty patients in working memory.
• They're moving between beds, capturing vitals on paper, asking questions across the room.
• They're juggling pending labs, pending consults, pending nursing tasks.
• At handover, they boil all of it down to two lines per patient — what matters now, what's pending — and read it out at 2 a.m.

There's no system in their workflow at all. There's an EHR they document into at the end. There's WhatsApp for the handover. The shift itself runs on paper and memory.

That gap — between the EHR record and the doctor's working memory during the shift — is where Medows lives.

Why "workspace" was the right word

We didn't want to build another record. The hospital already has a record, even if it's a flawed one, and rebuilding it is somebody else's company. What the doctor needed was a place to:

• See every patient on one card, color-coded by acuity.
• Capture vitals between beds in one tap.
• Ask the AI a question and have it already know which patient they're looking at — the lab values, the active orders, the tasks completed five minutes ago.
• Write the handover from the same context the AI used during the shift.

The word that kept fitting was workspace. Not a record. Not a search engine. A surface that holds the work while the doctor does it.

What we learned designing alongside residents

We've spent the last few months building with two residents — one at AIIMS Bhubaneswar, one at RG Kar Medical College — both doing their post-graduate training. Three lessons we keep relearning:

Tap targets matter more than features. A button that needs a precise tap on a smudged phone screen at 2 a.m. is a feature that doesn't exist. We tune touch areas more often than we ship new endpoints.

The AI has to be specific. Doctors don't want textbook prose. "Calcium gluconate 10 ml of 10% IV slow push" is a useful answer. "Treatment depends on the clinical context" is not. Every Medows answer cites a guideline (AIIMS, ICMR, NICE) so the doctor can verify and learn at the same time.

Handover is the killer app. The single moment with the most consequence in a shift is the handover. Doctors will adopt a tool that just makes handover better, regardless of everything else. So we made the handover almost involuntary — it composes itself as the shift happens.

What's next

Medows is free for individual clinicians while the product finds its shape. We're expanding from Indian pilots to UK, EU, and global English-speaking markets — particularly relevant after OpenEvidence withdrew from the UK and EU on 30 April 2026, leaving doctors there without a context-aware AI workspace.

If you're a doctor, try it: medows.ai. If you're an investor, hospital, or want to chat: alapanx@gmail.com.