Can Oncology Workflows Run Without Human Touch? - Anant Shankhdhar, Risa Labs
Trisca built a 4-agent AI pipeline automating oncology prior authorizations with zero human review
“confidence is a key metric that we were working towards”
Trisca deployed a four-agent system (EV, auth, necessity, submission) to fully automate cancer drug prior authorization workflows end-to-end, removing the mandatory human review step by using confidence scoring to identify orders safe for straight-through processing. The core engineering challenge is aggregating patient and insurance data scattered across dozens of portals and APIs. This is an early production example of agentic AI replacing human oversight in high-stakes clinical-administrative workflows, raising both efficiency and safety questions relevant to the broader healthcare AI space.