Anchr · Brainstorm
Walk in with a working formulation - not from memory.
Describe a focus and Anchr Guide reads the entire client record - mood, defusions, Living Compass, lesson notes, screening trajectory, Talk it Through transcripts, your signed notes and addendums - then returns a structured clinical brief. Not a generic AI summary. A genuine clinical artefact, built around the way a thoughtful colleague would orient you in ten minutes between cases.
Working formulation with required differential. Falsifiable techniques with literal opening lines. Stage-of-change calibrated. A learning loop that sharpens with this client over time. The prep you never had time for, every Tuesday morning.
- From "open" to "ready"
- ~10 min
- Frameworks Considered
- 9
- Claims Provenance-Tagged
- 100%
- Sharpened by you over time
- ∞
The entire chart - every field, every signal.
Brainstorm doesn't summarise the most recent note. It reads the chart end to end, the way you'd skim through it yourself if you had an extra hour before each session - and then it surfaces the patterns, the threads, and the absences you couldn't see in any single file.
Brainstorm synthesises across all of it - and returns one structured clinical brief.
Built on what a brief actually needs to be useful.
The Brainstorm output is structured around six clinical pillars. Each one exists because a generic AI brief wouldn't include it - and because the absence of any one of them is what makes generic AI outputs useless to you in practice.
Working formulation + required differential
The Guide picks the framework the data best supports - ACT (hexaflex), CBT, IPT, psychodynamic, DBT, schema, MI, CFT, integrative - and produces at least one alternative formulation, with what evidence in the record would adjudicate. Stops it (and you) committing to the first hypothesis.
Falsifiability on every recommendation
Each technique recommendation includes a "what would change my mind" line - the specific evidence in this client's data that would lower confidence in the technique fitting them. Without a falsifier, a recommendation is just an assertion.
Stage of change & phase of treatment
For each major theme, the Guide estimates the client's Prochaska stage (precontemplation → maintenance) and the phase of the therapy arc (early stabilisation, mid technique work, late integration). Recommendations are calibrated to where the client actually is - never pushing a precontemplative client toward action, never proposing schema work in session two.
Negative space - what's absent
The brief surfaces what should have come up given the data, and what hasn't. Chronic shame language without trauma history → trauma screen? Sleep dysregulation without substance-use mention → screen? Hopelessness language without a recent safety check? Surfaced before you walk in.
TIT-session divergence
The Guide compares what the client raised in their Talk it Through chats against what surfaced in the most recent signed note - and surfaces the gap. The thing they said to the AI but did not raise in the room. The 2am hopelessness that didn't appear in Friday's session. This is the comparison you literally cannot make alone.
Learning loop, calibrated by you
You rate which techniques landed and which didn't after each session. Future Brainstorms read those ratings: deepen what landed, retire what didn't, surface what changed if a previously-rejected technique might fit now. The artefact gets sharper for you, with this client, over time.
Scripts, not abstractions.
Every technique recommendation includes a literal opening line in quotes - what you could actually say to introduce it - plus a 2-3 step decision tree for what to do next depending on whether your client engages, hesitates or deflects, plus a minute estimate, plus what to deprioritise from the standard 50 minutes to make room.
You can implement a script. You cannot implement an abstraction. The brief is built so that even ten minutes before session, every line is something you can take into the room.
Opening line
- If they engage: walk through the "I'm having the thought that…" frame, hold ~6 min
- If they hesitate: drop the technique label, ask "what would help you not fight this thought right now?"
- If they deflect to content: let the content land for 2 min, return to defusion via embodied notice ("where does this thought sit in your body?")
Built for the way therapists actually think.
Beyond the six pillars and the literal opening lines, the brief has a handful of clinical disciplines baked into how it composes itself - each one a small but deliberate choice about what a useful brief looks like.
Lead with change, not steady state
If PHQ-9 or GAD-7 has crossed an MCID threshold (improvement or worsening) since the last brainstorm, the very first sentence of the brief leads with that delta - cited verbatim, computed in code, never generated by the AI. Clinically meaningful change always lands as a risk flag when the direction is wrong, never buried beneath techniques.
Cultural and contextual humility, made specific
The brief lists the specific, likely-relevant things Anchr cannot see for this client - cultural context, family system, identity, current life events, somatic presentation - rather than a generic disclaimer that applies to everyone. Calibrated humility, not boilerplate.
Self-of-therapist prompts
Two or three questions to sit with before walking into the session - colleague-to-colleague, never for the client. Examples of the right register: "What does this client activate in you that's worth noticing?" "Where might you over-identify, and what would that look like in the room?" "If you found yourself wanting to rescue or push, what would the data suggest the client is actually asking for?"
Provenance on every claim
The 2:14am Talk It Through, your note from April 18, defusion entry #14, the screener delta on Tuesday - every recommendation cites the field it was grounded in. You can audit exactly what each line was based on. Always your judgment. Never replaced.
Nine sections, in clinical order.
The brief composes itself in a deliberate clinical scaffold - not a chat answer. Risk content is surfaced before everything else; deltas before techniques; what's absent before what's present; the colleague-to-colleague self-of-therapist reflection always last.
Risk flags & deltas
MCID changes, crisis content, item-9 endorsements, safety-plan status - always first.
Working formulation
One framework selected, with phase indicator and one-line rationale grounded in data.
Required differential
At least one alternative formulation with what evidence would adjudicate.
Themes for this session
2-5 highest-signal patterns, each cited to mood entries, lesson notes, or TIT.
Techniques to trial
3-5 with literal opening lines, decision trees, time budgets, falsifiability.
Opening prompts
Cold-start questions that match the client's recent thread.
Unresolved threads
Topics from prior sessions never fully closed, with recommended re-entry.
What's absent / alternative lenses
What should have come up but hasn't. Other case conceptualisations to hold lightly.
For the therapist
Self-of-therapist prompts to sit with before walking in. Never for the client.
Sharpens with this client, over time.
Brainstorm is not a one-shot output. After each session you rate which techniques landed and which didn't. The next Brainstorm reads those ratings - deepening what landed, retiring what didn't, and surfacing if something previously rejected might fit now.
Brainstorm runs
Reads the chart end to end, returns the brief - formulation, techniques with falsifiers, what's absent.
You run the session
You use what fits, discard what doesn't, and find your own way through the hour - as you always have.
Calibration
Rate each technique: landed, partial, didn't. Next Brainstorm reads those ratings - the artefact sharpens with this client, every week.
Brainstorm never functions outside your oversight.
Brainstorm With Anchr Guide does not diagnose, does not prescribe, does not function as a therapist, and does not replace clinical judgment. The Guide's job is to sharpen your thinking, surface what you cannot see alone, and turn ten minutes of your time into the depth of context you've never had access to before. You hold the clinical authority - always.
Crisis content is surfaced, never silently filtered. Even low-confidence risk markers from the structured extraction propagate into the Brainstorm risk flags with explicit confidence and source annotation. If a client raised hopelessness in a Talk it Through chat that did not surface in the most recent session, that is the most clinically important divergence in the entire payload - and the brief leads with it.
Every claim is provenance-tagged. Every action is audit-trailed. AU-hosted, Privacy Act 1988, RLS-gated end to end. The audit trail is the artefact - you can always know exactly what the AI was reading when it produced a recommendation.
The prep you never had time for. Every Tuesday morning.
Ten minutes with Anchr Guide before each session and you walk in with the depth of context you've never had access to before - and the admin burden of catching up on a chart, lifted.
Anchr is an AI-assisted clinical platform built to complement therapy - never replace it. Brainstorm With Anchr Guide does not diagnose, prescribe, or function outside your oversight. Your clinical judgment is always the final authority.