
1. Technology overview
Soombit AI Inc. is located in Seocho-gu, Seoul, South Korea, and describes itself as about two years old. The company develops generative AI for radiology to reduce the time radiologists spend drafting reports.
AIRead-CXR drafts preliminary chest X-ray reports inside the reporting pane of the existing PACS. Radiologists review, edit and sign off each draft. It covers 57+ clinically validated findings on adult PA and AP radiographs, integrates with INFINITT, Taeyoung Soft and TECHHEIM PACS, and can run in the cloud or on-premise.
2. Highlighted features & specifications
- Draft reports: Generates preliminary chest X-ray reports for radiologist review and sign-off.
- Broad findings: Covers 57+ validated findings across lung, pleura, infection and TB, cardiac, mediastinal, bone and device categories.
- PACS integration: Works with INFINITT, Taeyoung Soft and TECHHEIM PACS, with other PACS supportable.
- Flexible deployment: Runs in the cloud or on-premise with installation by the vendor team.
- Regulatory scope: Approved for adult PA and AP chest radiographs.
- Roadmap: Prior comparison and per-hospital report formats are listed as coming soon.
3. Potential application
Hospital radiology departments can draft chest X-ray reports in PACS to shorten turnaround. Emergency departments can use faster preliminary reads to support triage decisions.
4. Benefits
Reported 42% reduction in reading time (34.2 seconds to 19.8 seconds, 5 readers, 758 radiographs). Thoracic radiologists rated 84.5% of drafts clinically acceptable in a 478-patient emergency study. Reported hallucination rate is 0.3%. Drafts appear in the existing PACS reporting window, so workflow changes are small.
5. Market trends & opportunities
Radiologist workload is growing faster than staffing, and vision-language models that draft reports are an emerging category. Vietnamese hospitals with high chest X-ray volumes, including TB screening, could evaluate such drafting tools, although the vendor lists no overseas approvals.
6. Other information
Based in Seoul. MFDS Class III approval no. 26-10 dated April 1, 2026, described as Korea's first generative AI medical device. Six papers in Radiology, AJR and European Radiology cited. AIRead-CT is listed for 2027 release.
One platform connecting startups, corporations, research institutions, regions and investors











Playbook & Handbook · February 6, 2026
Winning Accelerator Playbook 2026 by BambuUP
RETAIL E-COMMERCE: AI AND REAL-TIME DATA ARE TRANSFORMING HOW RETAIL BUSINESSES OPERATE








































