When Patients Started Talking, the Data Told a Different Story
Message response rates tripled in our first healthcare AI pilot. Then monitoring activity started falling at the same time, and the conversations explained why.
Read the article →Kestrel is running in a live clinical pilot, with real patients and real care teams. This is where we write about what that is teaching us, and about the ideas underneath it: why identity, memory and governance shouldn't belong to the model.
Message response rates tripled in our first healthcare AI pilot. Then monitoring activity started falling at the same time, and the conversations explained why.
Read the article →Message response rates tripled in our first healthcare AI pilot. Then monitoring activity started falling at the same time, and the conversations explained why.
Read article →A plain-language piece that defines the core problem Kestrel is solving and gives people a way to repeat the idea without sounding abstract.
Read article →Explains why long-term value in AI comes from retained context and why that context loses meaning if it lives inside someone else's platform boundary.
Read article →Frames the difference between prompting a model and governing an agent that is meant to persist, act, and hold trust over time.
Read article →