About

Let AUTOMATA write your reports — so you don't have to.

Built by a practicing school psychologist who got tired of retyping scores that were already on the PDF. Spreadsheet → formulas → code → app. Still tested on real caseload reports before anything ships. Nothing invented; every number traces to its source.

01

AUTOMATA

What we believe about a technical report.

AUTOMATA

makes

AUTOMATA

01

The thinking is the job. The typing isn't.

Interpreting a profile, weighing it against history and observation, deciding what it means for a student — that is professional judgment and it stays with the professional. Moving numbers from a score report into a table and from the table into sentences is not judgment. It is transcription, and transcription is where errors live.

02

A report is a legal document.

People sign their names to these. Eligibility decisions, services, and sometimes litigation rest on them. So a printed percentile from the publisher always beats a computed one, a descriptor always comes from the publisher's own band table, and a missing value is left blank and flagged rather than filled with something plausible.

03

Nothing runs without a yes.

Every automated check produces a proposal, not an action. You see what would change and approve it or reject it. Identifying data stays in your files. The system drafts on de-identified records and the report is re-assembled on your side.

02

Founder

School psychologist. Builder. He writes the reports this product was built for — and he got bored enough of writing them to build the thing that writes them for him.

PracticeSchool psychologist, Mira Monte High School, Kern High School District — special education assessment and eligibility

CredentialLicensed Educational Psychologist (LEP) · California Pupil Personnel Services (PPS) credential

EducationPhD, National University, JFK School of Psychology and Social Sciences — research on face perception and the cross-race effect

Contactinfo@drvalentinespsychlab.com

Kevin Valentine, PhD, LEP, is a school psychologist at Mira Monte High School (Kern High School District). His caseload is initials, triennials, re-evaluations, and the eligibility meetings that follow — each ending in a report built from score PDFs, rating scales, priors, and records.

He got bored of the typing, not the work. Retyping printed scores, rebuilding the same tables, re-explaining the same descriptors. So he taught himself the stack: spreadsheet shortcuts, then descriptor-band formulas (publisher printed value always wins), then PDF extractors, then an app. It was tested on reports he was about to sign. That pipeline is AUTOMATA.

His PhD (National University, JFK School of Psychology) was on face perception and the cross-race effect — different topic, same discipline: measurement, error, and what a number does and does not prove.

Why the "A"

The mark is an A drawn as a pen nib and a circuit trace: A for AUTOMATA, and the two things it joins together — the practitioner's hand and the machine that does the typing.

03

Student data

Where student data goes.

The short answer is that it stays with you. This is the long one.

01

It is read where it already lives.

Score reports, prior evaluations and rating scales are opened and read on your own computer, by software running on your machine. They are not uploaded to us. We do not receive them and we do not hold a copy, which means there is no account of ours for them to be exposed from.

02

The report is assembled on your side.

Tables, score graphs and the merged document are all built locally, against your own template, in your own folders. The finished report exists on your computer and wherever your district keeps it. It does not exist on our servers, because it was never sent to them.

03

A number we were never given cannot appear.

Every figure in a finished report traces back to the score sheet built from the source PDF. Anything that does not match is flagged rather than printed, and nothing is filled in because it looked plausible. That check runs no matter how the sentence around the number was drafted.

The rule underneath all three: identifying information has no reason to leave your machine, so it does not. If a feature ever needs outside help to phrase a sentence, what it will be handed is a de-identified summary — the area, and the finding your own software already computed — carrying no name, no date of birth, no school, no district and no identification number, with nothing retained once the sentence comes back. That is a design rule, not a setting you have to go and find.

04

How it got here

From one caseload to a company.

BoredomOne more report, one more evening retyping numbers that were already on a PDF. The realization that this was not what the job should be.

The spreadsheetEvery shortcut he could learn. Descriptor bands and percentile formulas for every instrument on the caseload, checked against the publishers' own tables, with a self-test that fails loudly when a band drifts.

CodeWhen formulas ran out, code. Score reports from WISC-V, WJ IV, WRAML-3, PH-3 (Piers-Harris 3), BASC-3, Vineland-3, and common rating scales read directly from the publisher's PDF into one structured data sheet, with the page each number came from.

AppsWhen code ran out, an app. Narrative, tables and score graphs composed from the data sheet against the practitioner's own template, every data point discussed in light of the suspected disability, with score-specific recommendations tied to eligibility.

The verifierEvery claim checked before the practitioner sees it. Mismatches flagged, never fixed silently. Nothing overwrites a report without a yes.

AUTOMATAThe same pipeline, productized and offered to other psychologists, clinicians and evaluators — and next, to engineering, compliance, laboratory and research reporting, where the problem is identical.

Try the live demo See plans