Afini.ai
Sign in
← Back to use cases

Therapists who want an instrument, not a replacement

Big Five with norms from a million respondents, AAP, AHP, ATP. Comparable data your patient can bring you done. No AI telling you how to work.

We do not sell pseudoscience and we do not substitute clinical training. We sell a quantified cognitive profile that the client brings to session — with public-domain or published instruments, clear norms and documented limitations. If you find it useful, you use it. If not, you don’t.

TerapeutaPacientePerfil cognitivoportable · JSONentre sesiones

The pains you bring

Expensive, impractical psychometric tests

NEO-PI-R requires a licence, a manual, scoring, country norms. The patient takes it in office or at a centre. Slow. Expensive. IPIP-NEO is public domain, normed, validated against NEO-PI-R with correlations of .77 to .86 (Goldberg, Johnson). Your patient completes it at home for €5–20.

Patients arriving with pseudo internet tests

"I am INFJ", "I am an enneagram 4", "I am an alpha wolf". You explain MBTI has terrible test-retest reliability (Pittenger 1993, Boyle 1995) and the enneagram has no serious empirical base. Better the other way round: patient arrives with real data, and if they want, you help them interpret it.

General-purpose AIs offering "psychotherapy" without training

Replika, Pi, Character.ai promise "your best friend" or "your coach". Your client tells serious things to a chatbot with no clinical supervision and no referral protocol. Afini does not play that game: our AI is calibrated to your profile, but the whole system reminds you it is not a therapist and recommends finding one when warranted.

What Afini solves

One page with percentiles and facets

The test.afini.ai report delivers percentiles per domain (N, E, O, A, C) and across all 30 facets, normed against 1,461,383 respondents (Johnson IPIP-NEO-300, IPIP-NEO-120, Open Psychometrics). Your patient brings it as a PDF; you read it in five minutes.

Attachment layer (AAP) and humour layer (AHP)

On afini.ai you add the AAP (Afini Attachment Profile, Bilbao AI S.L., 2026) for anxiety, avoidance and disorganization, and the AHP — Afini Humor Profile (Bilbao AI S.L., 2026) — for the six humour dimensions: playful-affiliative, reflective-resilient, witty-creative, corrective-satirical, dark-transgressive and ridicule-vulnerability. Data comparable to what you would use in your own assessment, without the patient going elsewhere.

ExAfiniTwin for your own analysis

If the patient purchases the AfiniTwin (249€), they can share it with you as JSON. You read it with any tool. You have raw, normed data with an interpretation protocol, no middleware. You cross it with your own theoretical frame (psychodynamic, CBT, systemic, whatever you use).

We do not interfere with your work

Afini does not diagnose, does not interpret clinically, does not recommend treatment. It gives data. Where the data ends and clinical reading begins is your seat. The personalised AI, if used, has explicit instructions to refer to a professional at signs of pathology or crisis.

A sample conversation

The patient brings their profile. You read it.
My patient brought the Afini PDF. N4 = 84, A6 = 22. 38 years old, professional gridlock, comes in for "I can’t decide anything". What stands out?
High N4 (social shame) + low A6 (firmness) is an unkind mix: heavy fear of judgement, low capacity to assert. The hypothesis isn’t indecision — it’s that any decision triggers anticipated shame. Before asking him to "decide something", help him tolerate the idea of being wrong in public. Different work.
And how do I verify the hypothesis in session?
Ask about a small recent decision he did make. If he says "none", you have the confirmation: the memory bias rewrites even the trivial. If he names one, look at how long he spends justifying it — low A6 people justify each choice as if expecting a reproach.

A typical run-through

A week using Afini in clinical practice

  1. First session of a new patient

    Patient arrives with the test.afini.ai PDF you requested at booking. You save 30 minutes of biased self-report: you already have normative baseline. You open the interview through the most extreme dimensions, not "tell me what brings you here".

  2. Hard case between sessions

    You can’t stop turning over a patient’s case. You paste her profile (anonymised) into your AI and ask for alternative hypotheses. Without the profile you’d get 5 generic ones; with it, three are discarded and two deepened. Useful so you don’t obsess.

  3. Closing a process

    A patient who finished six months ago retakes the test. You compare percentiles: N3 down 12 points, A6 up 9. Not magic, it’s a measurement; but it gives her something concrete to say when asked "are you better?". And it confirms to you the work landed where it should.

What you are probably asking

What evidence is there of IPIP-NEO ↔ NEO-PI-R equivalence?
Goldberg (1999) reported correlations of .77 to .86 between IPIP and commercial scales. Johnson (2014) published norms for the 300-item version with 145,388 respondents (Personality and Individual Differences). IPIP facets correlate robustly with their NEO-PI-R counterparts, though they are not identical.
What about GDPR and patient data?
The patient owns their profile. Individual account, exportable as JSON, deletable on request. They share it with you by their own means. Bilbao AI S.L. does not train models on user data. Legal detail at /legal and /privacy in six languages.
Can I recommend it to patients without joining myself?
Yes. test.afini.ai needs no signup, just the link. If you want to understand the product before recommending it, the logical step is to take it yourself. Otherwise, patients can do it and bring you the PDF with no involvement on your end.

Your case, with your own data

Start my profile

Cancelable, exportable, no training on your data.

Keep exploring

Afini for therapists: validated instruments at the service of your clinical work | Afini.ai