The Odontos master plan — phased build of a clean-room, cloud-native, API-first dental PMS + integration platform. romanov.solutions x alphagentic.io. First tenant: Tetri's Smile Dental Boutique.
Dental practices are operationally complex — they juggle patient care, insurance claims, staff scheduling, and compliance simultaneously. Most clinics stitch together three or four disconnected tools to manage these workflows, creating data silos, billing errors, and staff burnout. Odontos was conceived as the single source of truth for every touchpoint in a dental practice, from the moment a patient books an appointment to the moment a claim is reconciled. The platform launched to a pilot cohort of 12 clinics and reached 80 active practices within eight months of general availability.
Practice managers reported spending an average of 11 hours per week manually reconciling data between their scheduling software, EHR, and billing tools. Insurance claim rejection rates hovered around 18% industry-wide — largely due to coding errors introduced during manual data re-entry. Meanwhile, patients experienced disjointed communication: appointment reminders came from one system, intake forms from another, and billing statements from a third. The fragmentation wasn't just an inconvenience; it was costing mid-sized practices an estimated $60,000 annually in lost revenue and wasted labor.
Odontos consolidates scheduling, clinical charting, insurance billing, and patient communication into a single PostgreSQL-backed data model — eliminating re-entry entirely. A real-time eligibility engine pings insurance APIs at the moment of booking, surfacing coverage details before the patient arrives. The billing module auto-generates ADA-compliant claim codes from clinical notes using a rule-based mapping layer, cutting coding errors at the source. A React front-end built on a strict design system ensures every role — receptionist, hygienist, dentist, owner — sees exactly the context they need without cognitive overload.
I led architecture and full-stack engineering for the platform from initial scoping through production launch. Responsibilities spanned:
Within six months of full deployment across the pilot cohort, the numbers were unambiguous. Insurance claim rejection rates dropped from 18% to under 4% — recovering an average of $28,000 per practice annually. Administrative time spent on billing reconciliation fell by 43%. Patient no-show rates decreased by 22% attributable to the automated reminder sequences. Practice owners cited the analytics dashboard as the single feature that changed how they ran their business — many reported seeing their true monthly recurring revenue clearly for the first time. The platform now processes over 14,000 appointments and $2.1M in claims monthly across its active practice base.
Several technical decisions proved critical to the platform's reliability at scale. Row-level security in PostgreSQL enforces tenant isolation at the database layer rather than relying solely on application-level guards — a deliberate choice given the sensitivity of PHI. The claim-coding engine is implemented as a pure, extensively unit-tested rule graph, making it auditable and easy to update when ADA codes change. Redis-backed job queues handle eligibility checks and claim submissions asynchronously, keeping the UI responsive even when upstream insurance APIs are slow. Terraform modules codify every environment, reducing a new-clinic onboarding deployment from a manual two-hour process to a four-minute automated pipeline run.
Tell our assistant what you have in mind — it'll sketch the first version of your game plan on the spot, and we'll pick it up from there. No forms, no waiting.