THE METHOD

Weeks of preparation, not years of search.

Therapy works. It also takes time to know what you're there for. TwentyThird compresses the preparation — the part where you learn what you're actually saying — so the room itself can go faster.

THE PIPELINE

Five stages, one brief.

From sampled speech to a working document a clinician can read in eight minutes.

sampling volume across the cohortFig. 01
STAGE 01

Sampling

Voice and text from guided sessions. The model listens for shape, not content.

words sampled~24,000
sessions18
elapsed~9 wks
median turn42 wds
extraction feature classesFig. 02
STAGE 02

Linguistic extraction

Features the model lifts from the transcript — quietly, before any interpretation.

01qualifier density
02modal verbs
03conditional frame
04affective hedges
05slips · misnamings
06repetition windows
recurrence three anchors of one patternFig. 03
STAGE 03

Recurrence detection

The pattern that returns at the same shape, in different decades. The signal that something is structural.

  • 2014first jobceiling at month 9
  • 2019secondceiling at month 11
  • 2024thirdceiling at month 8
reconstruction working draftFig. 04
STAGE 04

Script reconstruction

The sentence underneath the pattern. Old line, then a candidate revision the subject can rehearse.

01 · I amnot yet ready.already allowed.
02 · towait to be chosen.name what I want.
clinical handoff working briefFig. 05
STAGE 05

Map for the clinician

Three labelled meters and a one-line structural reading. Eight minutes to read, weeks to use.

case 0217-Bprepared 04 · 26
Ego structure.72
Object relations.48
Symbolic register.67
Reading
obsessional · with hysterical traces
SCOPE

What we measure, and what we refuse to.

What we measure
language patternsQualifier density, modal verbs, hedges, conditional frames.
recurrencePatterns that return at the same shape across years.
narrative coherenceHow a life is told, where the seams show.
attachment markersWhere intimacy crosses a threshold, what follows.
linguistic slipsMisnamings, condensations, displacements.
What we don't
DSM / ICD categoriesNosology is a separate instrument, used by clinicians.
IQ or cognitionOut of scope. The model is psychodynamic, not psychometric.
personality typesFive-factor and its descendants are not part of the brief.
risk scoresAcute risk is a clinician's call, not a model's.
predictionsStructure, not forecast. We name what is, not what comes.
Compression

Walk in knowing what to say.

Therapy is not the place to figure out the question. It is the place to live with it. The compression we offer is on the approach — the months a clinician usually spends listening for the shape of the brief.

Across the pilot cohort, prepared subjects reached working alliance in fewer sessions. The reading was already there.

alliance median sessions to working allianceFig. 06
Time to working alliancemedian across cohort
1714unprepared23prepared
day 1day 7day 14day 23
~6fewer sessions to working alliance
CONTINUE

Begin, or view a sample report.

Begin →or view a sample report