River
Y CombinatorBacked by Y Combinator
FREE TEMPLATE

Compliance Training Documentation Template

Three documents and three sheets that name which of three evidence forms, a certification, an LMS transcript, or a timestamped report, each requirement actually demands.

Free download  ·  No account needed

Evidence Register

Programme:

One row per requirement. The verdict comes from opening the artifact on file, never from a delivery system's status.

CellRequirementEvidence FormMatchPartialWrong FormExpiredMissing
R-01
R-02
R-03
R-04
R-05
R-06

Totals

RowWhat it countsValue
T-1Lines currently due
T-2Marked complete by the delivery system
T-3True match, evidence form checked
T-4Read complete but fail the form check

The gap between row T-2 and row T-3 has no line here until every artifact on file is actually opened and checked against the form its own rule names.

Search compliance training documentation and every guide converges on one checklist: an immutable timestamp, a verified learner identity, a content version, a tamper-evident audit trail. All four are good advice, and all four assume every requirement wants the same record shape. They do not. A rule that says an employer shall certify wants a signed document with named fields. A rule stated in hours wants proof of duration. A rule about one hazard wants content tied to that hazard. None of the three is a timestamp with better metadata attached to it.

Marrowfield Cold Storage runs a refrigerated warehouse in Fresno, California and one in Amarillo, Texas. Six role, topic and jurisdiction combinations sit on its Requirement Register: forklift operators, refrigeration maintenance technicians, an annual facility inspection, and California's own supervisory and non-supervisory harassment-prevention training, plus hazard communication for anhydrous ammonia. That is 135 register lines, four not yet due inside a new hire's own grace period, 131 currently due. Delivery-system dashboards report 119 of those 131 complete, 90.8 percent. Amarillo has no equivalent harassment-training mandate at all, so its staff never appear on those two rows.

Reading the actual artifact behind each of those 131 lines finds only 104 true matches, 79.4 percent. Six forklift certifications are missing their evaluation half. The annual facility inspection is filed as a maintenance log rather than a certification, and eight hazard-communication sessions are filed as a paper sign-in sheet rather than the content-versioned transcript the rule names. All fifteen read as finished in the dashboard that drives most training programmes. Once a job task analysis decides training is mandated, this is the register that proves the record behind it is real.

119 read complete. 104 the evidence actually supports.

The Requirement Register, the Completion Tracker's own rate against the Evidence Register's true match rate, and the lines flagged partial, wrong form or expired.

Requirement Register

Six requirements, 135 register lines, three evidence forms

Marrowfield Cold Storage, a two-facility refrigerated warehouse operator. One row per role, topic and jurisdiction, never one row per employee.

CellRole or SubjectJurisdictionLinesCadenceEvidence Form
R-01Forklift OperatorsFederal25Eval. every 3 yrsCertification: training + eval date
R-02Refrigeration Maint. TechsFederal9Retrain on changeCertification: name + date
R-03Ammonia System (facility)Federal2AnnualCertification: equipment + inspector
R-04Fresno SupervisorsCalifornia6Every 2 yrsTimestamped report: 2 hr min
R-05Fresno Non-SupervisoryCalifornia39Every 2 yrsTimestamped report: 1 hr min
R-06Warehouse Ops StaffFederal54Initial + new hazardLMS transcript: course + version
Total2 facilities1353 forms

Amarillo, Texas carries no statutory harassment-prevention mandate, so its staff are not on rows R-04 or R-05 at all. Six rows, never one row per employee.

Completion Tracker vs. Evidence Register

119 read complete. 104 the evidence actually supports.

The same 131 currently-due lines, read two ways.

CellRole or SubjectDueSystem CompleteEvidence MatchGap
R-01Forklift Operators2522166
R-02Maintenance Techs9770
R-03Ammonia System2211
R-04Fresno Supervisors6440
R-05Fresno Non-Supervisory3530300
R-06Warehouse Ops Staff5454468
Total13111910415

90.8 percent by the dashboard, 79.4 percent by the register. R-02, R-04 and R-05 lose nothing in the gap, their overdue lines are already correctly flagged there too. The 15-line difference sits entirely in R-01, R-03 and R-06.

Flagged Lines

What each gap line actually is

Every flagged line names the specific missing field or the specific wrong artifact, never a percentage.

CellRequirementVerdictOn FileRequired
FL-07R-01 ForkliftPartialTraining date, evaluator blankEvaluation date + evaluator identity
FL-02R-01 ForkliftExpiredLast evaluation 2023-02-14Re-evaluated every 3 years
MT-05R-02 Maint. TechMissingNo record foundCertification, hired 2026-06-01
AMA-FACR-03 Ammonia SystemWrong FormMaintenance log entryCertification naming equipment + inspector
SUP-04R-04 SupervisorExpired2-hr session, 2023-11-02Every 2 yrs, window closed 2025-11-02
HC-19R-06 Warehouse OpsWrong FormPaper sign-in sheetLMS transcript: course + version

Six verdicts, never a paraphrase: match, partial, wrong form, expired, missing, not yet due. A wrong-form line and a missing line both score zero on a completion percentage; only one of them has anything on file at all.

What's in the pack

01

Evidence Register

One row per requirement, cross-checked against the actual certification, transcript or report on file, with six verdicts: match, partial, wrong form, expired, missing, and not yet due.

02

Requirement Register

One row per role, topic and jurisdiction, each carrying its own citation, its own cadence and trigger, and which of three evidence forms the rule itself names.

03

Completion Tracker

What a normal delivery system already reports, with due dates computed from the register's own cadence rather than a platform's default reminder window.

04

Training Content

The outline for each topic pulled from what the standard's own training clause requires, with the practical evaluation noted wherever the rule pairs one with the classroom half.

05

Attestation Format

A blank specimen of each evidence form actually in use, with the specific fields the underlying rule names rather than a generic name-and-date template.

06

Evidence Standard

How a verdict gets assigned, the retention floor once a record exists, and who a wrong-form or expired line routes to.

How to use it

  1. 1

    Open it in River, or download it

    Open the pack in River and have the agent build your own register and run the check, or download the six blank Word and CSV files with no account.

  2. 2

    Send the citations and the roster

    The specific standards or statutes that actually apply, your roster by role and location, and whatever completion data your current delivery system already exports.

  3. 3

    Get the register and the forms

    One row per role, topic and jurisdiction, each naming its evidence form and its cadence, with a blank specimen of every form in use.

  4. 4

    Run the check against what's on file

    Every completion opened and read against the form its own row demands, verdicted match, partial, wrong form, expired, missing or not yet due.

Frequently asked questions

Is this template free?

Yes. Download all six files as Word documents and CSV sheets with no signup and no card. Edit with AI is the optional path, where the agent reads your own citations and checks what is on file against the form each one names. The rest of the packs sit in the template library.

What format are the downloaded files?

Word documents (.docx) for the training content, the attestation format and the evidence standard, and CSV (.csv) for the three register sheets, zipped together. They open in Word, Pages, Google Docs, Excel, Numbers and Sheets. In River the same content opens as live Docs and Sheets.

What's the actual difference between a certification and a certificate?

A certificate is what a learner downloads at the end of a module and keeps for themselves. A certification is the employer's own signed record, and OSHA's forklift standard names exactly what belongs on it: the training date, the evaluation date, and the identity of whoever performed each. A generic template built for a simpler rule quietly drops the evaluation half.

Does the register decide what the training itself has to cover?

It names the citation and the evidence form each row demands, and the content outline is pulled from what the standard's own training clause requires rather than a generic vendor course on the same topic. OSHA's hazard communication rule names detection methods and the hazards actually present as required content, and a module that never names the real chemical does not satisfy it.

How is this different from what our LMS already reports?

An LMS reports whether an assignment was opened and finished, which is a real fact and not the same fact a regulator asks for. This register opens the artifact behind that status. On the worked example, a dashboard reporting 90.8 percent complete drops to 79.4 percent once every certification, transcript and report is actually read.

Does this replace a training programme's design and evaluation plan?

No. Training Program Pack designs the modules, the minutes and the evaluation levels once training is already the answer. This pack starts after that decision is made and asks a narrower question: whether the record behind each completion is the evidence form the requirement's own citation names. Whether training changed anything is separate, and Training Evaluation and Impact Pack answers that instead.

The training content only exists in one expert's head. Where does that fit?

Before this. Send whatever is already delivered along with the completion export, and River builds the register and runs the check against what's on file. If the content itself only exists in one expert's head, start from the interview before there is anything to check evidence against.

Find out which of your completions the evidence actually supports

Download the blank pack as Word and CSV files, or open it in River and have the check run against your own citations, roster and completion data.

Edit with AI