Start with the same core questions
Regardless of size, confirm the intended workflow, client population, privacy model, provider controls, accessibility, support, data portability, review timing, and urgent-support routes. Ask for evidence that matches the claimed use.
Priorities for a solo practice
A solo provider usually needs fast setup, low administrative overhead, clear client onboarding, and a single view of work awaiting review. Complexity that cannot be delegated becomes a direct burden.
- Can one provider configure and maintain it?
- Are common actions fast enough for daily use?
- Can the provider control reminders and sharing?
- Are pricing and cancellation terms predictable?
- Is support available without an IT team?
Priorities for a group practice
Groups add role permissions, consistency, onboarding, supervision, reporting, and staff turnover. The best tool for an individual clinician may become difficult when each person configures it differently.
- Role-based access and least-privilege defaults
- Shared templates with local provider control
- Practice-level adoption and workflow reporting
- Documented onboarding and offboarding
- Clear ownership of alerts and review queues
Priorities for IOP, PHP, and coordinated programs
Programs may need daily cadence, cross-disciplinary visibility, transitions between levels of care, and explicit escalation workflows. Evaluate capacity under a realistic caseload and test what happens during weekends, staff absences, and discharge.
Run a pilot that can answer a decision
Choose a representative group, establish a baseline, define success and guardrail metrics, and set a review date. Measure activation, useful engagement, provider time, review timeliness, client burden, support issues, and privacy or safety events.
A concise buying checklist
- What exact client and provider problem will this replace or improve?
- Which team owns configuration, consent, and review?
- Can we export our records in a usable format?
- Which security and contractual documents can we review?
- How are AI features disclosed, controlled, and changed?
- What implementation help and response times are included?
- What will we stop doing if we adopt this?