AI should help the doctor, not replace the doctor
AI should sort the inbox so the doctor can decide. Here is where we draw the line.
Thyra Blog
Rule-based and AI-powered decision support integrated into clinical workflows.
AI should sort the inbox so the doctor can decide. Here is where we draw the line.
A clinical look at RAG-based tools that let clinicians ask natural-language questions across a patient's entire chart, and what separates real longitudinal search from note-search with a chat box.
HTI-1 forces certified EHRs to disclose how predictive AI tools are built, trained, and validated. Here's what it actually requires, and what it doesn't.
AI chart audits scan for HCC coding gaps, unsupported diagnoses, and documentation-to-billing mismatches at a scale manual review cannot match. Here is what they actually catch.
AI safety in healthcare is now a workflow control problem. Here is what healthcare IT leaders should verify in audit logs, approval workflows, and source grounding before deploying an AI EHR.
AI hallucinations in clinical workflows pose significant patient safety risks. Essential prevention strategies include data grounding, deterministic rules, and audit trails.
AI scribes assist with documentation. An EHR overlay addresses selected workflows alongside an existing system. Compare their scope and confirm supported connections.
Outpatient EHRs optimized the note. The inbox became the channel for non-visit care, and it lacks a safety-grade triage layer.
How AI-native EHR workflows with Smart Inbox, Smart Search, and SMART on FHIR overlay are transforming endocrinology and primary care. Built by an endocrinologist for specialty and primary care practices.
A clinical inbox is not email. It is a task engine with safety requirements. Here is how to design one that routes, owns, and resolves work deterministically.
CGM data is abundant. Clinical time is not. A scalable CGM workflow needs automated summaries, trend comparison, and defensible documentation in one screen.
Prior auth is the most hated workflow in healthcare. Automation helps, but only if it addresses the real bottlenecks: data assembly, payer rules, and follow-up tracking.
Not all AI scribes are equal. A maturity model helps clinics evaluate what they are actually buying and what level of trust it deserves.
Keep exploring