AI Philosophy
AI Should Support Clinical Judgment, Not Replace It
Compass exists to help providers think, not to think for them.
Artificial intelligence has enormous potential to improve behavioral health care.
Its role, however, should not be to make clinical decisions.
Leitara was designed around a different philosophy: AI should reduce administrative burden, organize clinical information, and surface patterns for provider review while preserving the provider's judgment as the final authority.
Compass exists to help providers think, not to think for them.
The Human Remains in Control
Every recommendation generated by Compass follows a review-first workflow.
Rather than modifying records automatically, Compass prepares structured proposals that can be reviewed, edited, accepted, or rejected before any changes are applied.
This ensures that AI becomes part of the clinical workflow without becoming the clinical decision maker.
Review-first principles
AI suggests
Providers review
Providers edit
Providers approve
Providers remain accountable
Clinical judgment always belongs to the provider.
Context Matters
Large language models are most effective when they understand context.
Compass brings together information already present within the therapeutic relationship, including:
Client Activity
- Homework
- Journals
- Check-ins
Clinical State
- Goals
- Topics
- Prior Compass
- Provider notes
Structured Context
Client activity and clinical state are combined and organized into meaningful clinical context.
Compass
Transforms structured context into actionable, reviewable suggestions for the provider.
Rather than generating isolated responses, Compass reasons across structured clinical information to produce recommendations grounded in the client's ongoing therapeutic journey.
Structured Output Over Open-Ended Generation
Free-form conversations are useful.
Structured recommendations are actionable.
Compass transforms clinical information into proposed updates for existing workflow objects rather than producing long narrative responses that require interpretation.
Examples include
Treatment goal updates
Homework recommendations
Session topic suggestions
Provider follow-up tasks
Resource recommendations
Check-in assignments
By producing structured recommendations instead of free-form content, Compass remains transparent, reviewable, and easy to incorporate into clinical workflows.
Designed for Trust
Trust is built through transparency.
Every recommendation is presented alongside the information used to generate it, allowing providers to understand why a suggestion was made before deciding whether it should become part of the clinical record.
This review-first model encourages collaboration between clinician and AI rather than automation without oversight.
Built Around Safety
Compass was intentionally designed to operate within defined clinical boundaries.
Compass never
- Diagnoses
- Communicates independently
- Modifies records
- Operates outside authorization
Instead, Compass assists with organizing information that providers have already collected while preserving existing authorization, audit, and review workflows.
A Foundation for the Future
As behavioral health platforms continue to evolve, structured clinical engagement will become increasingly valuable.
Homework, journals, longitudinal outcome measures, treatment goals, and client reflections create a growing body of clinical context that can support increasingly capable AI systems.
Leitara was designed with that future in mind.
By combining continuous client engagement with structured provider review, Compass establishes a foundation for AI-assisted care that remains transparent, explainable, and clinically responsible.
By reducing cognitive load, organizing meaningful clinical context, and supporting provider decision making, AI allows clinicians to spend less time managing information and more time focused on the person in front of them.
That is the role of Compass.
And that is the future Leitara is building toward.
Explore the Platform
Learn more about the principles, architecture, and technology behind Leitara.