01/06/2026
Our Cosylabers, Bor Brecelj and Tim Soles Kos, returned from the Mayo Clinic AI & Data Science Datathon and Data Summit 2026 in Jacksonville with fresh insights from an intense few days of workshops, talks, and discussions around real-world clinical AI deployment.
Tim shared what stood out to him most:
“It wasn’t the size of the dataset or even the fact that our team won a prize. It was realizing how large the gap still is between clinical AI concepts and the software hospitals actually use in daily workflows. I went to Mayo without a specific goal in mind, but after three days of lectures and conversations, I came back with a much clearer picture of where clinical software is still falling short, and who is quietly working to fix it.
Prof. Kee Yuan Ngiam, a physician from Singapore, presented BERTRAND, a multi-agent AI system connected to Epic that generates pre-rounding and pre-clerking drafts from clinical notes and results.
Dr. Chris Sauer showed a similar direction from Essen, where integrated data infrastructure supports clinical workflow tools, including bedside dashboards for handovers and AI integration.
Both are physician-engineers building these systems themselves because the tools they need in practice still don’t really exist. It’s a strong signal that clinical insight and software engineering need to evolve much more closely together.”
Thank you to the Mayo Clinic team and the course directors for organizing an excellent event and creating space for these conversations to happen.
If you’d like to continue the discussion, you can meet Tim at PTCOG 2026.
