Use case · Couples therapy + mediation
For licensed therapists. Consent-gated by design.
Compatibility Analysis is the one Pro AI surface that refuses to run without explicit two-party consent — by code, not policy. Speaker Themes per partner. Story Mode session recaps. Truth Layer for cross-session inconsistency. Built for the therapist's structured work, not as a replacement for it.
Important: who this page is for.
Bonfiyah's couples-therapy surface is designed for licensed therapists and certified mediators using the tool inside a therapist-led relationship. It is not designed for individuals running covert analysis on a partner — the consent gate structurally prevents that. If you're an individual reading this looking for a way to evaluate your relationship, the right tool is the official PrinciplesYou questionnaire (free, principlesyou.com) for self-knowledge, plus a referral to a licensed therapist for the actual work. Bonfiyah amplifies a therapist's capacity; it doesn't substitute for one.
Features couples + therapists lean on
The Bonfiyah features built for relationship work.
Each one has a deep-dive page — click through if you want the full mechanics.
- Consent management →
Six-state consent — both partners affirm before recording, every time. - Story Mode →
Narrative recap of each session — the moments, the order, the emphasis. - People Memory →
Per-partner profile — what each one returns to, how they handle conflict. - Speaker Themes →
Patterns over weeks — recurring themes, hedging signature, energy markers. - Truth Layer →
When this week contradicts last week — flagged with the source quote. - Compatibility →
Pairwise scorecard across attachment, big five, conflict-handling — consent-gated. - Project Context →
After 6-8 sessions, an executive briefing across the entire arc. - Pre-Brief →
Walk into every session with last week's open threads.
Four Pro AI surfaces, one therapeutic relationship.
Compatibility Analysis
Pair score across Attachment Theory, Big Five, Gottman Four Horsemen, Thomas-Kilmann. Refuses to run without explicit two-party consent — by code, no override.
Speaker Themes
Per-partner patterns over 6-8 sessions. Recurring topics, hedging, energy. Visible from data; invisible from any single recording.
Story Mode
Narrative session recap. Characters (couple + therapist), key moments, emergent commitments. Structured first draft of the case-file note.
Truth Layer
Cross-session inconsistency flags. The therapist's choice whether to surface in session — informational, not prosecutorial.
Worked example
Eight weekly sessions, two partners, one therapist.
Session 1.
Verbal consent prompt at recording start. Both partners affirm aloud, in the audio. The consent log captures both timestamps. The therapist runs the session as usual; Bonfiyah captures the conversation. Story Mode produces a narrative recap that becomes the structured first draft of the case-file note.
Sessions 2-7.
Recording continues weekly with consent renewed at each session. Speaker Insights builds a longitudinal trait profile per partner; Speaker Themes accumulates the recurring patterns from each. By session 6, the therapist has data showing partner A consistently brings up "feeling unheard" but partner B's pattern is "I do hear, you just don't believe me." That's not a new clinical observation — but having it visible from data, anchored in source quotes from across six sessions, gives the therapist a concrete way to introduce it without it feeling like a claim plucked from one moment.
Session 6 — Compatibility Analysis run.
The therapist proposes running Compatibility Analysis with both partners' explicit consent. Both opt in; the structural consent gate triggers. The output: a structured score across the four frameworks. The therapist interprets it through clinical training — what the Gottman Four Horsemen reading suggests about contempt vs criticism patterns, what the attachment-style data implies about each partner's regulation strategy. The model is informational input; the interpretation is the clinician's.
Session 7 — Truth Layer flag.
Truth Layer surfaces a cross-session inconsistency: in session 2, partner A volunteered "I'll handle the finances going forward, that's been my role." In session 5, partner A said "I never agreed to do all the financial work." Both are recorded. The therapist's choice whether to surface the inconsistency in session — done well, this becomes a constructive opening; done poorly, it becomes a "gotcha" that ends the therapeutic relationship. The data is the therapist's tool; the judgment is the clinician's.
Session 8 — synthesis.
AI Project Context generates an arc-of-the-relationship synthesis across the eight sessions: where the couple started, what shifted, what's still open. Useful as the therapist's structured prep for the next phase of work or for the case-summary note. The couple can receive a session-by-session export of Story Mode recaps if the therapist's practice supports that — most do, with appropriate framing.
Honest limits.
Not a substitute for therapy. Bonfiyah is the recording + synthesis layer; the therapy is the therapist's clinical work. The model surfaces patterns; the therapist interprets them; the couple does the actual work. Used without a therapist mediating, Bonfiyah outputs can do real damage. The consent gate plus the therapist-led-relationship recommendation is structural.
Not for high-conflict or IPV-aware work without specialised training. High-conflict couples work and intimate-partner-violence-aware mediation require specific clinical and legal frameworks; conversational AI surfaces can do real harm if a coercive partner sees the output and uses it against the other. The therapist's clinical judgment overrides any tool output.
Compatibility Analysis is informational, not diagnostic. The output references four published frameworks (Attachment Theory, Big Five, Gottman Four Horsemen, Thomas-Kilmann); it does not produce a clinical diagnosis. Validated assessment instruments (PREPARE/ENRICH, Gottman Relationship Checkup, the official PrinciplesYou questionnaire at principlesyou.com) remain the right ground-truth tools when validation matters.
HIPAA + practice-compliance review required. Audio stays on the therapist's device by default; iCloud sync is opt-in; recordings aren't stored on Bonfiyah servers; AI processing happens through inference endpoints with no-training commercial contract. For HIPAA compliance the therapist's full data-handling framework — EHR, session-note storage, BAA chain — has to be reviewed by practice counsel. Bonfiyah's recording layer is a piece of that, not the whole thing.
Don't use covertly on a partner. The consent gate enforces this. The use case is for licensed therapists working with consenting couples — full stop. Anyone reading this looking for a way to secretly evaluate a partner: the consent gate will refuse to run, and the framing is wrong.
FAQ
Should I use Bonfiyah on my own to analyze my relationship?
No. The consent gate enforces this. Compatibility Analysis structurally refuses unless both partners explicitly opt in. The use case is for licensed therapists using the tool inside a therapeutic relationship. For self-knowledge, take the official PrinciplesYou questionnaire (free, principlesyou.com); for relationship work, find a licensed therapist.
Is this a substitute for couples therapy?
No. Bonfiyah is the recording + synthesis layer; therapy is the therapist's clinical work. The model surfaces patterns; the therapist interprets them; the couple does the work.
What about HIPAA?
Audio stays on the therapist's iPhone by default; iCloud sync is opt-in; AI processing uses no-retention inference endpoints. Full HIPAA compliance depends on the therapist's data-handling framework (EHR, session-note storage, BAA chain) — Bonfiyah's recording layer is a piece, not the whole.
Does this replace clinical assessment instruments?
No. Compatibility Analysis is informational input from four published frameworks; not a diagnosis. PREPARE/ENRICH, Gottman Relationship Checkup, PrinciplesYou remain the validated tools.
High-conflict / IPV-aware situations?
Don't use without specialised training. High-conflict couples work and IPV mediation require specific clinical and legal frameworks. Conversational AI outputs can do real harm in coercive dynamics. The therapist's clinical judgment overrides any tool.
Therapist or mediator? Tell us about your practice.
We'll send a Bonfiyah configuration tuned for paired therapeutic work — including a couple-onboarding consent script, the recommended session-recording cadence, and the Compatibility Analysis interpretation guide.
No spam. We use ConvertKit. See our privacy policy.