The technical detail, in one place.
For technology directors and practice IT teams: how QuillsFlow is built, how it connects, and how AI and data stay governed. Capabilities are pilot-stage or planned unless stated otherwise.
How the complete PIMS and the integration layer share one governed core.
Do both deployment paths use the same platform?
Yes. The complete QuillsFlow PIMS and the integration layer share the same claims logic, analytics, workflow engine and audit controls. Only the deployment differs: the PIMS captures data natively, while the integration layer maps records from an existing PIMS.
How is a claim modelled?
A claim is an explicit state machine, from estimated through submitted to settled, so every status change is traceable. A fee engine shows how amounts are derived, and a double-entry ledger keeps money movement reconcilable.
How is clinic data isolated?
Each clinic's data sits inside a tenant boundary with its own isolation, encryption and retention rules, so one practice's records never inform another practice's outputs.
How Quill connects with practice systems and insurers.
Which PIMS connections are planned?
Planned connectors include Cornerstone, ezyVet and Vetspire. Each connector is validated per PIMS during the pilot before it is offered more widely.
Which insurer exchange formats are supported?
Partner-specific REST, EDI, FHIR, HL7 or secure file (SFTP) interfaces map into a canonical claim workflow, with validation and traceable status events. Transport and field mappings are chosen for each implementation.
Can outside AI assistants connect?
Quill offers permission-scoped MCP tools that expose approved actions and context to outside assistants, without unrestricted access to clinical systems.
How AI outputs stay reviewable and under clinician control.
How does the AI scribe work?
Ambient capture drafts structured SOAP notes mapped to record fields. Every draft stays unsigned until the clinician reviews and signs it.
How are claim-readiness checks explained?
Rules and model signals flag missing documentation, coding conflicts and likely submission gaps, each with a plain-language reason for staff review. They support clinic decisions and do not determine coverage.
What happens when the AI is unsure?
Hybrid retrieval requires every answer to cite a source. When confidence or citations fall short, the system withholds an answer instead of guessing.
The technical layers behind the pet owner experience.
Data foundation
Longitudinal pet identity; clinic records; labs; medications; owner-reported outcomes; wearable and claim events.
Analytics
Personal baselines, time-series change, missing-data awareness, multi-signal correlation, and population comparison where appropriate.
Predictive modeling
Condition-risk signals and review tiers that retain contributing features, timing, and uncertainty for inspection.
Operational intelligence
Clinician-governed rules convert insights into routing, messages, tasks, reminders, and measurable completion.
Trust layer
Permission-aware access, provenance, explainability, human review, and an auditable record of decisions and actions.
Controls designed in from the start.
How is access controlled?
Role-scoped identity, segmented networks and permission-aware data access limit each user to what their role needs.
Is there an audit trail?
Tamper-evident audit records capture who accessed what and when, including automated recommendations and the human decisions made on them.
What is the SOC 2 status?
A SOC 2 audit is planned. The Trust Center describes current controls and what is still in progress.