About
Built for the people whose decisions decide everything.
Most "AI for X" products write summaries. We score people. There's a difference — and for therapists, recruiters, lawyers, founders, mediators, and the rest, it's the only difference that matters.
Why we built Profyls
We watched generic LLMs swallow every category — except the ones where being wrong matters. A therapist needs to know a relapse spiral is coming. A hiring manager needs to know the rehearsed answer from the real one. A litigator needs to predict how a judge will rule, not how a chatbot would summarize the brief.
Generic AI guesses. Psychometric instruments measure. We chose measurement — composed from the same peer-reviewed frameworks academic psychology has trusted for decades, applied to the conversations that actually decide things.
189 metrics. 8 frameworks. 9 verticals. One platform. No black boxes.
Backed by exceptional minds
A working sample of our advisor bench (composite names shown for privacy):
Hon. R. Marquez (Ret.)
Former federal magistrate · judicial profile advisor
Dr. K. Hollis, PhD
Clinical psychology · language & affect
M. Okafor, PhD
Computational linguistics · LIWC contributor
S. Park
GP, Series A fund · GTM advisor
J. Reyes, JD
Litigation partner, AmLaw 100
Our principles
Four lines we will not cross.
Theory before output.
Every Profyls metric maps to a peer-reviewed framework. We never ship a number we can't cite.
Your data is the work, not the product.
Zero data training. No cross-customer signals. Delete-on-demand. Always.
The professional decides.
Profyls is an instrument, not a verdict. The clinician, lawyer, founder, or coach reads — and decides.
Compliance is a vertical, not a checkbox.
HIPAA for clinicians. ABA Rule 1.6 for lawyers. EEOC for hiring. GDPR/CCPA everywhere.
The road here
A short history.
2022
First clinical pilot
Three private-practice therapists, twelve clients, eight weeks. Drop-out prediction beat clinician gut by 2.4×.
2023
Frameworks unified
Licensed LIWC-22 + 7 peer-reviewed instruments. Built the composition layer that became Profyls.
2024
Vertical playbooks ship
Lawyers, recruiters, mediators, and founders join the first cohort. Adverse-impact audits clear.
2025
189 metrics. 9 verticals.
SOC 2 Type II. HIPAA-BAA available. Today: scoring high-stakes language at scale, with the receipts.
Where we say no
The customers we turn away.
A short list — but a real one. Profyls is built for measurement, not manipulation.
Predictive policing or carceral scoring.
We will not deploy psychometric scoring into criminal-justice triage, parole, or sentencing pipelines. Full stop.
Covert relationship surveillance.
No spousal monitoring, no stalkerware integrations, no romance-scam adjacent use cases. Consent is the floor.
Employee-monitoring as performance management.
Risk and compliance workflows? Yes — with audit and disclosure. Wall-to-wall keystroke psychology? No.
Political profiling at the individual level.
Cohort-level psychographics for civic messaging is in scope. Individual partisan scoring is not.
Who built this
Three rooms. One product.
We're a small team — by design. Every shipped metric passes through all three rooms before it sees a customer.
The lab
Computational psycholinguists who've spent careers inside LIWC, SALLEE, Plutchik, and Schwartz. They keep us honest about what a score does — and doesn't — mean.
The practice
Clinicians, litigators, mediators, recruiters, and founders embedded in product. Every workflow is co-designed with the people who'll use it on a Tuesday morning.
The build
Engineers who've shipped at Stripe, Anthropic, Palantir, and Two Sigma. Latency budgets, audit logs, and SOC 2 are first-class citizens — not afterthoughts.
What's next
The 2026 roadmap, briefly.
11 languages, full parity.
Receptiviti's multilingual LIWC extends to Spanish, German, French, Portuguese, Italian, Dutch, Arabic, Russian, Mandarin, and Japanese. We bring vertical playbooks to all of them.
Real-time scoring on live audio.
Sub-200ms streaming scores during Zoom, Teams, and phone calls — for negotiations, depositions, and clinical sessions where the next turn matters.
Customer-owned models.
Bring-your-own baseline corpora for Firm tier — fine-tune drives, values, and emotion calibration to your population, with full auditability.