Imaging Models
Some of the clinical evidence that prior authorization depends on doesn't exist as structured data — it exists as an image. A scan, read by a clinician, that contains the measurements and findings a payer's policy actually asks for. Today that evidence gets transcribed by hand. The imaging-models layer is what we're building to change that.
OphthaLogicIn development
We're building OphthaLogic, an imaging model designed to read ophthalmic scans — starting with OCT — and extract the clinical evidence a payer's policy requires directly from the image. It's designed to remove the manual step where someone reads findings off device software and retypes them into a submission. OphthaLogic is in development; it is not yet part of a live submission.
Beyond OCT
OCT is the starting point because it's where ophthalmology's anti-VEGF evidence lives, but the underlying problem isn't specific to one scan type. Over time, we expect the imaging-models layer to extend to other modalities, in other specialties, wherever the clinical evidence that decides a prior authorization lives inside an image rather than a form field. That's a direction we're building toward, not a dated roadmap.