Defensible restructuring

A RIF plan that survives the deposition.

Most restructuring analysis lives in a spreadsheet that was edited forty times by six people. When it's challenged, nobody can say which version the decision was made on. OrgTool replaces it with a workbench whose output is deterministic: same inputs, same result, same hash — every time, on any machine.

See it live Sample audit pack
01
Draw the selection
Multi-select the candidate pool from your live org. Every click re-runs the analysis.
02
Watch the law evaluate
EEOC 80% rule with Fisher exact and 2-SD tests. Federal, California, New York, New Jersey, and Illinois WARN thresholds by establishment. Cost impact, annualised.
03
Export the record
One ZIP: report.json, selection.csv, adverse-impact.csv, README.txt, and cover.pdf — sealed with a SHA-256 of the inputs. Recompute it any time to prove nothing changed.
FOR INFORMATIONAL PLANNING PURPOSES ONLY. NOT LEGAL ADVICE. OrgTool computes the tests; qualified employment counsel makes the call.
What counsel receives
report.json
Full analysis, input hash, timestamps
selection.csv
Roster with demographic codes, salary
adverse-impact.csv
Rates, ratios, p-values, z-scores
cover.pdf
Cover sheet with the input-hash block
README.txt
Provenance, disclaimer, verification steps

Two tests, one deterministic ratio.

The EEOC's Uniform Guidelines (29 C.F.R. § 1607.4(D)) set the four-fifths rule as the primary screen: a selected group's rate must clear 80% of the highest-rate group's, or adverse impact is presumed. OrgTool runs that rule against every demographic field you've coded, plus Fisher's exact test as a corroborating statistical-significance check, and an optional 2-SD test.

This is a statistical screen, not a legal conclusion. A failed test means the selection needs a documented, defensible business justification — that judgment call belongs to counsel, not to software.

WARN, jurisdiction by jurisdiction.

Federal WARN and the state statutes with materially different thresholds — California, New York, New Jersey, and Illinois — are evaluated against your selection's headcount and establishment. OrgTool doesn't claim coverage of every state; it evaluates the five it has rulepacks for and says so, rather than silently guessing at the rest.

Rulepack: Federal · California · New York · New Jersey · Illinois

See the workbench on your own org.

Open OrgTool
FAQ

Frequently asked questions

Answers restate claims verified file-by-file against the current build — see claims/ in the site repo. Last verified July 2026.

What is the EEOC 80% (four-fifths) rule?
Under the Uniform Guidelines on Employee Selection Procedures (29 C.F.R. § 1607.4(D)), a selection rate for any group that is less than 80% of the highest group's rate is generally regarded as evidence of adverse impact. OrgTool runs that screen against every demographic field you've coded, with Fisher's exact test as a corroborating statistical-significance check and an optional two-standard-deviation test.
Is OrgTool giving legal advice?
No — and the product won't pretend otherwise. Every restructuring output carries a non-removable banner: OrgTool computes the tests; qualified employment counsel makes the call. A failed screen means the selection needs a documented, defensible business justification — that judgment belongs to counsel, not software.
Which WARN laws does OrgTool evaluate?
Federal WARN plus the four state statutes with materially different thresholds in the shipped rulepack: California, New York, New Jersey, and Illinois. OrgTool evaluates the five jurisdictions it has rules for and says so, rather than silently guessing at the rest.
What exactly does counsel receive?
A single ZIP: report.json (full analysis, input hash, timestamps), selection.csv (roster with demographic codes and salary), adverse-impact.csv (rates, ratios, p-values, z-scores), cover.pdf (cover sheet with the input-hash block), and README.txt (provenance, disclaimer, verification steps).
What does the SHA-256 hash actually prove?
That nothing moved after the fact. The pack is deterministic — same inputs reproduce a byte-identical archive with the same hash, on any machine, any time. On Enterprise, a countersignature additionally proves that this input hash was submitted by a licensed installation at a stated time. What it does not prove: that the roster was accurate or the decision lawful — those remain the customer's.
Can a regulator or opposing counsel verify a pack without OrgTool's cooperation?
Yes. Our Ed25519 public key is published on the security page; verification runs offline against the pack's .sig.json, or through an unauthenticated API endpoint. No account with us is required — by design.
Does any of the RIF data leave my machine?
No. The analysis is computed locally and the pack exists only where you save it and send it. There is no vendor-side copy for a breach, a subpoena, or a curious admin to find.
What happens when a selection fails the screen?
You see it before anyone else does. The failure appears live as you draw the selection, so you can redraw, watch it re-evaluate, and involve counsel while the plan is still a plan — not a filed exhibit.
In one paragraph

OrgTool's Enterprise restructuring workbench turns a RIF plan into a defensible record: draw the selection from your live org and the EEOC four-fifths (80%) rule, Fisher's exact test, an optional 2-SD test, and federal + California/New York/New Jersey/Illinois WARN thresholds evaluate on every click — locally, with nothing uploaded. The export is a deterministic audit pack (report.json, selection.csv, adverse-impact.csv, cover.pdf, README.txt) sealed with a SHA-256 of its inputs, so the same inputs reproduce a byte-identical archive months later and counsel — or opposing counsel — can verify its integrity without an account. It computes the tests; it is not legal advice.

Copy-paste friendly — for the email to whoever decides.