AUTOMATAby Valentine Labs

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

Upload the score-report PDFs and the student's records. That's it. AUTOMATA builds the tables, graphs the scores, discusses every data point as it relates to the student's disability — or lack of one — and writes score-specific clinical recommendations based on the eligibility. You review it, and sign.

Runs on de-identified records. Names and identifiers never leave your files.

01
Feature tour

What happens between "upload" and "sign."

You do the first step and the last one. AUTOMATA does everything in between, and every step leaves a trail.

01

Upload — the PDFs and the records, as they are

Drop in the score reports (WISC-V, WJ IV, WRAML-3, PH-3, BASC-3, Vineland-3, rating scales), prior evaluations and student records. AUTOMATA reads them directly; you never retype a number. Every score, percentile, confidence interval, date and descriptor lands in one structured data sheet with the page it came from. A printed value always wins over a computed one.

PH-3 Score Report.pdf (sample) → data sheet TOT T 38 %ile 10 p.2 BEH T 45 %ile 81 p.2 printed wins INT T 40 %ile 17 p.2 PH DATE 12/14/2025 p.1, column match 9 scales · 52 tags · 0 dropped
02

Descriptors — from the publisher's own band table

"Low Average" means what Pearson, Riverside or WPS says it means, for that instrument. Every band table in AUTOMATA is checked against the publisher and covered by a self-test that fails loudly if a cutoff ever drifts. Tails read "<1" and ">99", the way the report prints them.

WRAML-3 GMI 68 → Extremely Low (7-band table) WJ IV "<40" → <1 · Very Low (bound kept) WISC-V VSI 84 → Low Average self-test: 150 / 150 checks pass
03

Write — tables, graphs, and every data point in context

AUTOMATA builds the score tables and graphs, then discusses each data point as it relates to the student's suspected disability — or the absence of one — in your own template and voice. Strengths and needs, previous-assessment summaries and the eligibility reasoning are written the same way.

Working memory was an area of significant need. The Working Memory Index of 79 (8th percentile, 95% CI 73–88) fell in the Very Low range — a processing weakness consistent with the suspected Specific Learning Disability+ score table · + index graph · every bold value linked to its cell
04

Recommend — score-specific, tied to the eligibility

Clinical recommendations are generated from the actual scores and the eligibility category, not from a generic list: a Very Low working memory index under an SLD finding produces different recommendations than the same score under OHI, and an Average one produces none. Prior testing is carried forward into a score-history table alongside.

WMI 79 · Very Low · SLD (auditory processing / memory) chunk multi-step directions; provide written copies preferential seating; check for understanding extended time on tasks with heavy memory load each recommendation cites the score that produced it
05

Verify — every claim, before you see it

Every score in the prose is matched to the table. Every descriptor is matched to the band. Section order, required headings and formatting are audited against your standard. A mismatch is flagged, never silently fixed. A missing score leaves the sentence blank and flagged — AUTOMATA never fills a gap with a plausible number.

27 of 27 cited values traced to a cell 5 of 5 descriptors match band table ! PSI missing from table — sentence left blank, flagged heading order matches template
06

Approve — nothing runs without a yes

The audit is a proposal, not an action. You see what would change, accept or reject each item, and sign. Identifying data was never in the draft: AUTOMATA works on de-identified records, and the finished report is re-assembled on your side.

3 proposed fixes · you choose [ accept ] descriptor "Low" → "Low Average" [ reject ] reorder "Background" above "Observations" [ accept ] add PH DATE to summary table nothing written until you click
02
Under the hood

Built for the report you actually sign.

InstrumentsWISC-V, WJ IV, WRAML-3, PH-3, BASC-3, Vineland-3, rating scales — more on request
InputsPublisher score PDFs, spreadsheets, scoring portals, prior evaluations, records
OutputTables, score graphs, narrative and recommendations in your template — Word or Google Docs
PrivacyDe-identified drafting; identifiers stay in your files; nothing identifying stored
RecommendationsScore-specific and eligibility-based; each one cites the score behind it
PercentilesPrinted wins; formula is the fallback; "<1" / ">99" tails preserved
DescriptorsPublisher band tables, self-tested; never invented
VerificationNumber-to-cell, descriptor-to-band, format-to-template; flags, never silent fixes
ControlEvery automated change is a proposal you accept or reject
03
Plans

Pay for the evenings you get back.

Every plan is the full pipeline on your own templates. Cancel any time. Your reports and your data are always yours.

USD.
01 / solo

Practitioner

$49 / month

One school psychologist, clinician, or evaluator. Unlimited reports.

  • Psychoeducational & clinical templates — or bring your own
  • Extract from publisher PDFs, spreadsheets and portals
  • Compose in your voice; verify every number
  • Prior-testing comparisons carried forward
Start Practitioner
02 / team

Practice

$149 / month

Up to five evaluators sharing templates, style rules and a caseload.

  • Everything in Practitioner, for five seats
  • Shared template library and house-style rules
  • Caseload sync: one data sheet feeds every report
  • Audit log of every check and every approval
Start Practice
03 / district

District & Enterprise

Custom

Unlimited seats, single sign-on, your compliance review, your paper.

  • District templates, eligibility criteria and forms built in
  • SIS and Google Workspace integration
  • DPA, FERPA/HIPAA documentation, security review
  • Onboarding and training for your assessment team
Talk to us

Full plan details and FAQ on the pricing page. Every plan starts with a live walkthrough on one of your own reports.

Get started

Bring us your worst report. We'll show you AUTOMATA write it.

Tell us what you write, what you write it from, and how long it takes today. We'll reply from a real inbox.

Opens a pre-filled email to info@valentine-labs.com — no account, no mailing list.