Building an AI Tool for Therapists Without Recording Sessions (and why that matters)

Over the past few months, I’ve been building an AI tool called AfterSession that helps therapists generate structured notes after sessions.

One thing that surprised me early on was how divided the space is around recording sessions.

Many tools record therapy sessions and then generate notes from transcripts. On paper, that sounds ideal — but when I started talking to therapists, I kept hearing the same concerns:
• Privacy concerns around recording
• Clients feeling uncomfortable being recorded
• Compliance complexity
• Workflow friction

So instead of building around recording, we focused on something simpler: brief summaries.

Therapists can speak or type a short summary after a session, and AfterSession turns that into structured notes (SOAP, DAP, BIRP, etc.).

This approach ended up solving a few unexpected problems:
• No session recording required
• Faster workflow after sessions
• More clinician control over the final note
• Lower privacy concerns

We’re still early, but a few things I’ve learned so far:
1. Small workflow improvements matter more than big AI features
2. Privacy concerns are much bigger in healthcare than I expected
3. Therapists care more about speed and control than “AI magic”

I’m curious — for those building AI tools in regulated industries, have you found that “simpler AI” often works better than more ambitious approaches?

Would love to hear how others are thinking about this.

If anyone is curious, this is what we’re building:
https://aftersession.ai