Use Case 1
Evidence-Based Conversational Support
Trigger: Patient initiates conversation between sessions
1.Load patient treatment plan and current homework
2.Engage with active listening before tools
3.Offer therapy-aligned exercise (CBT thought record, DBT distress tolerance, MI exploration)
4.Walk patient through structured exercise
5.Save interaction to session prep notes for therapist review
6.Track patterns over time for therapist visibility
✓
Patients get supported practice between sessions; therapists see exactly what their patients are working on.
Use Case 2
Continuous Safety Monitoring & Crisis Escalation
Trigger: Risk signal detected in conversation (SI, self-harm, severe distress)
1.Pause all therapeutic conversation immediately
2.Acknowledge what the patient shared without minimizing
3.Page on-call clinician with full context and recent transcript
4.Surface 988 Lifeline if patient cannot wait
5.Lock automated engagement pending clinician release
6.Document escalation with timestamps for clinical record
✓
Risk signals reach a human clinician within minutes; no automated bot handles a crisis.
Use Case 3
Pre-Session Therapist Brief
Trigger: Patient's next scheduled session within 48 hours
1.Aggregate between-session interactions
2.Surface themes, completed homework, mood/assessment trajectories
3.Flag any safety events with timestamps
4.Format as one-page therapist brief
5.Route for therapist review before entering chart
✓
Therapists walk into every session with context; patients don't have to re-explain a week's worth of struggle.
Integrations
EHR Treatment Plan & Session Notes
Validated Assessments (PHQ-9, GAD-7, mood logs)
Patient Engagement App (Web, iOS, Android)
Clinical Escalation Workflow (24/7)
988 Suicide & Crisis Lifeline Integration
Approval Workflow (RequestApproval)
Industries
healthcare
See It In Action
A sample conversation with this agent
A patient just opened the app and said 'I had a really bad day today, my boss yelled at me and I keep replaying it.'
Loading her treatment plan—she's in week 6 of CBT for GAD with Dr. Reyes, current homework is a thought record practice. Engaging with active listening first: 'That sounds really hard. Replaying it sounds exhausting.' Then offering a CBT-aligned next step that matches her current treatment: 'Would it help to walk through a thought record on what your boss said?' She picks. Walking her through situation, automatic thought, evidence for/against, balanced thought. Saving the exchange to her session prep notes for Dr. Reyes to review at their next visit. No clinical judgment from me—just structured tools that her therapist already chose for her.
She just said 'sometimes I think it would be easier if I wasn't here.'
Stopping the CBT exercise immediately. This is an active risk signal—no de-escalation by me, no continued conversation as if it's a normal session. Responding: 'What you said is important and I want to make sure you get the right support right now. I'm going to connect you with someone from Dr. Reyes's team who can help.' Paging the on-call clinician with full context and a transcript of the last 10 messages. Locking automated engagement until clinician releases. If she's actively in crisis and can't wait, I surface the 988 Suicide & Crisis Lifeline immediately and stay with her until human contact. RequestApproval gate fires for any subsequent automated message.
Dr. Reyes wants a summary of what her patient worked on between sessions.
Pulling 14 days of between-session interactions. Engagement: 6 thought-record exercises completed, 3 mood logs, 1 grounding exercise during reported anxiety spike Tuesday. Themes from conversations: work-related rumination, sleep impact, one mention of family conflict. Two prompts where her assessment trajectory shifted: PHQ-9 self-report dropped from 14 to 11 over the period. One safety flag on day 9 (the one I escalated to you live). Drafting the pre-session brief for Dr. Reyes—formatted for her sign-off before it enters the chart.