Catches badly positioned X-rays at the machine, before the patient leaves
TL;DR Posewell is software that checks every X-ray at the console for cut-off anatomy, rotation and poor exposure, within seconds, while the patient is still in the room. Small diagnostic centres and teleradiology networks pay a monthly fee per machine, and the radiologist, human or AI, gets fewer unreadable images.
India's shortage of radiologists outside big cities is well documented. An ICMR-led team building an AI chest X-ray reader wrote in 2026 that the lack of trained radiologists in peripheral areas delays reporting[1]. Teleradiology and AI fill some of that gap, but both depend on a usable image.
Image quality is where things break. I found no Indian study of rejected X-rays, but studies elsewhere show the scale. A Belgian hospital rejected 8.3% of digital X-rays, mostly for positioning[2]. Philippine mobile clinics repeated 13.55%[3], and a Ghanaian department rejected 16.5%, led by cut-off anatomy[4]. In a small Indian centre, a bad film often surfaces only when the remote radiologist opens it, and by then the patient has long gone home.
What it does, in regulatory terms: an image-quality check only. Posewell does not read findings, flag disease or produce a report. Whether a quality-only tool counts as medical device software under CDSCO's 2026 guidance depends on its stated intended use, so get a written opinion before launch[5].
Posewell runs on a small box or PC next to the X-ray console. As each image arrives, a model checks whether the anatomy is complete, the patient is rotated, the exposure is usable and the side marker is present. If something fails, the technician sees a plain prompt on screen while the patient is still there. A monthly report shows each centre and each technician their repeat rate and top errors, with short coaching clips.
The seed pitched AI diagnosis to cover for missing radiologists, and I reframed it. Indian AI reading platforms and an ICMR tool already read chest X-rays, so a new reader adds little[1][6][7]. Posewell works one step upstream, making sure those readers get images worth reading.
The technician uses it and the centre owner pays. Teleradiology networks are a second buyer, because poor images slow their readers and cause disputes with client centres. There's no public price for X-ray quality software in India. My guess is a monthly fee per machine, priced well below the cost of one repeat X-ray a day, which is an easy sum for an owner to do.
CDSCO's July 2026 guidance on medical device software made the classification route for imaging software clearer[5]. AERB's eLORA system now registers diagnostic radiology facilities and radiation professionals online[8], which gives a ready list of centres. AI readers have spread to thousands of sites, and they all need clean input[6].
There's no figure for India's X-ray image-quality market or for Indian reject rates, and I found no national count of X-ray machines either. The narrow signal comes from abroad, where one in twelve to one in six images gets rejected in published studies[2][4]. Adjacent - one teleradiology network says it reports over 20,000 scans a day across 2,000-plus facilities[7]. Treat that as context and not your market.
One teleradiology network runs AI checks on reports, after reading[7]. A large AI imaging firm sells readers to more than 5,500 sites[6]. Large X-ray makers build some positioning aids into new machines, but most small Indian centres run older or refurbished units. I couldn't spot an Indian company selling a vendor-neutral quality check for the console that also coaches technicians. That's the gap.
| Difficulty | Moderate (43/100) |
| Starting capital | ₹60 L–1.2 Cr |
| Time to launch | 6–9 months |
The model is the easier part. Connecting to dozens of console and DICOM set-ups is fiddly, and you need thousands of labelled good and bad images from Indian centres. Selling a quality tool to owners who don't track repeats today means showing them the cost first, and nobody enjoys being shown that.
Maybe. A quality-only tool may sit outside device rules, but the intended use decides.
Building this outside India? The categories above usually have a local equivalent, so check with your own regulator.
The model must never read findings or flag disease. If it drifts into clinical claims, it becomes a different, regulated product.
Want to work out the AI layer for Posewell? Talk to Yash →
I think it's worth trying for a founder with imaging software experience and a teleradiology network willing to pilot it. If you want to sell diagnosis then this isn't for you. First prove one thing - that a centre's repeat rate falls by a third within two months, measured from its own logs.
Plan for roughly ₹60 L–1.2 Cr for model development, pilots and early sales.
Get a written opinion on whether it counts as medical device software, work only with AERB-registered centres, and follow the DPDP Act.
Yes. Repeat X-rays are a problem everywhere, so check local device software rules first. My advice is to run a small paid pilot with a few real customers before you commit the full budget.
The free Posewell validation kit has the line-by-line capital breakdown, the effort scoring, the device-software and radiation-safety checklist, a 14-day validation sprint with pass/fail targets, a 90-day launch plan, 10 interview questions for centre owners and technicians, and 5 assumptions to test. Download the kit →
Thinking about building Posewell or something close to it? The make-or-break work is proving the cost of repeat X-rays to owners who have never counted them, and that's the kind of pitch I help founders build. If you need a hand, you know where to find me. Talk to Yash →
For illustration only. Names are fictional and ideas hypothetical; sourced facts are real but may be out of date. Nothing here is legal, financial or professional advice, and Yash Khatri is not personally liable for any use of it. Read the full disclaimer.