I dreamed about it, and now a startup has built it ! 👇
(recently acquired by ) just announced a voice transcription AI module aimed at improving obstetrics ultrasound reporting & workflows, and this is a big deal ! 👇
I still remember my first ultrasound shifts when I started practicing as a radiologist in private practice. It wasn’t easy and it still feels a bit painful today. You see a high volume of patients, back-to-back in different rooms. After the exams, you head back to your workstation to report — sometimes hours later, at the end of the day. You’re reviewing a small sample of images from the full exam, trying to recall what you saw earlier in the ultrasound room, and then dictating the report. It’s inefficient and prone to errors.
At the time, I thought, “I need to dictate the report immediately after the exam, while still in the room.” But the room wasn’t equipped with a workstation, and I didn’t have a good solution. I even considered using my iPhone or another recording device, but quickly abandoned the idea — it would just add more friction to the workflow.
What Sonio claims to solve is exactly this issue. Now, while performing the ultrasound, I can comment on my findings and immediately generate a report using their AI voice solution. This will ensure I don’t forget anything or mix up information from different patients. I haven’t tested the tool yet, but if it also integrates with AI image analysis to include automated measurements or other AI outputs, it could make the reporting even more comprehensive, and much easier and faster, radically reducing the repetitive manual entries.
Voice is a central part of how physicians interact with their tools and environment, so integrating this kind of input into AI software is a crucial step for adoption and success.
One word of caution, though — voice reporting during an exam, especially with the patient in the room, could unintentionally disclose worrisome findings too early. Imagine talking to your AI system with the patient right there: “5 cm mass suspected of a tumor.” So, it’s important to communicate findings at the right moment, and with care.
I didn’t expect this innovation to come from an ultrasound AI startup. Traditionally, speech transcription has been led by companies like Nuance (Microsoft) or innovative startups dedicated to voice analysis like , or . Now, I can picture a future where sonographers focus on moving the transducer on the patient with minimal clicks and keyboard interaction. Instead, they would communicate directly with the AI system, then quickly generate and validate the report before they even leave the ultrasound room. There’s no need to review images separately or start the report from scratch. It’s a game changer for both quality and productivity.
Voice-driven workflows are a promising area of innovation in radiology and medicine, and I’m excited to follow where this trend takes us ! 🚀
Link to press release :
