Building a Custom Care Management Platform For a US Based Company

We built tenant management as the first layer, envisioning the platform as a scalable solution that could eventually serve other non-profits and healthcare providers. Each tenant operates in isolation, with its own users, configurations, and data, which was essential for privacy and compliance.

We implemented 9 distinct roles mapped to specific stages of the operational workflow, from intake coordinators and care managers to billing managers and physicians. Every role has tightly scoped permissions, ensuring staff access only to what is relevant to their work and that sensitive actions are restricted to the right reviewers.

Patient management was the heart of the platform, split into two clearly defined stages to ensure explicit handoffs and clear accountability.

Intake coordinators handle onboarding through a structured form paired with an in-app interaction recorder. Consent is captured before recording begins, resolving a long-standing problem in which intake conversations were lost or summarized inconsistently. Once onboarded, the client profile moves through a defined lifecycle, with a Patient Journey view providing any authorized user with a time-stamped history of every event from the assessment onward.

After onboarding, the patient is assigned to a care manager who owns the assessment stage. The system supports MCP-specific assessment flows: forms, required fields, and downstream billing rules adapt automatically based on the payer. This eliminated one of the biggest sources of rejected claims: using the wrong documentation template for the wrong payer.

Billing was the most operationally painful area, and we focused the most engineering effort there. We built a configurable billing engine that defines rules per MCP and applies them automatically when claims are generated. Billing teams work through structured queues mirroring the claim lifecycle, with pre-billing readiness checks that flag missing documentation or mismatched codes before a claim is ever sent.

The platform connects to a clearing house partner to support the full claim settlement lifecycle, from initial submission and eligibility verification through denials, resubmissions, and remittance processing. Every billing action is logged with timestamps and user attribution, giving the team an audit-ready trail for internal reviews and compliance audits.

Two AI capabilities were core to the platform: transcription of recorded client interactions and assistance with care plan generation. Both were initially scoped against third-party APIs, but at the volumes that the company was projecting, third-party costs would have grown into a meaningful operating expense over time.

We proposed a self-hosted approach featuring  Whisper for speech-to-text. Deployed on the client’s infrastructure, giving the team accurate transcription of intake and assessment conversations without per-minute API charges.

We also used Gemini to draft care plans and integrated it into a controlled flow in which care managers receive a structured, suggested plan based on assessment data, which they can review, edit, and finalize.

Before the platform, generating any report meant pulling data from spreadsheets by hand. We built a reporting layer giving leadership real-time visibility into clients onboarded, care plans active, and claims submitted, approved, and rejected by MCP. Dashboards are customizable for funder-specific reporting requirements, which vary significantly in a non-profit context.