Outsourced EmploymentPowered by StealthAgents

Practical guide

Run a training completion audit with source evidence

A training completion audit that distinguishes attendance from completion, ties each record to a source, and fixes the system instead of the report.

Published

Practical guide illustration for Run a training completion audit with source evidence
Table of contents

Define completion before counting it

Write a plain definition of completion for each course or requirement. The definition may include a final assessment, a signed acknowledgement, or a practical demonstration. A login alone is not completion, and an attendance sheet is not proof of understanding. The definition should match the rule that made the training necessary.

Separate required training from optional training before building the sample. A required course has a population, a deadline, and a consequence for a missing record. An optional course may still need a record, but it should not inflate the completion rate. Mixing the two hides the gap that matters.

Record the version of the course that the employee completed. A course that changed its content may need a new completion date. The old record can still be valid for the old version, but it cannot prove the new requirement. The audit should compare like versions.

Map each requirement to its source system

List the system that holds the authoritative completion record for each requirement. A learning platform, a signed form, a supervisor observation, and a certificate can all be valid sources when the policy names them. Record the system and the field that proves completion. A report built from a secondary system should point back to the primary source.

Check whether the source system captures the employee identifier used by HR. A course record with a name and no identifier can be matched to the wrong person. Record the identifier and the match method. If the systems cannot be joined, the audit should say so rather than guess.

Include the date the source was last refreshed. A completion report pulled before the deadline may be accurate for that moment and stale by the review date. Record the export time and the reporting window. The reader should know what the number does and does not cover.

Sample the population before the deadline

Freeze the required population at a stated date and include new hires who joined during the window. Record the source of the population and any manual additions. Do not remove a difficult case because the record is hard to find. A replaced case should be reported as a deviation.

Take a sample across roles, locations, and managers rather than only the largest group. A single team may have a working habit that the rest of the organization does not share. Record the sample size and the selection method. The sample should be large enough to show a pattern without becoming a full re-audit.

Review the sample before the compliance deadline so the team can act on the finding. An audit completed after the deadline produces a history lesson rather than a correction. Set the review date with the course owner and keep it visible. The deadline for the audit should be earlier than the deadline for the training.

Trace each completion to its evidence

For each person in the sample, open the source record and confirm that it matches the requirement. Record the completion date, the course version, and the evidence location. A screenshot of a dashboard is weaker than an export that shows the identifier and the date. Prefer the source that the policy names.

Record a missing source as missing, with the reason and the date it was checked. Do not infer completion from a calendar invitation or a manager's memory. A missing record may mean the training did not happen, or it may mean the record was not captured. The audit should distinguish the two before assigning a repair.

Ask the course owner to review any case where the requirement and the record disagree. The owner can confirm whether an equivalent course satisfies the requirement. Record the decision and the date. The audit should not invent an equivalence that the policy does not allow.

Fix the source instead of the report

Group the findings by cause and look for a system problem before blaming the employee. A course that does not send completion data to HR will produce missing records no matter how many reminders are sent. A roster that excludes new hires will always undercount. Record the cause and the owner who can change it.

Assign one fix with a testable result, such as a nightly export that includes the employee identifier. Set a date to confirm the fix with the same sample method. If the gap remains, inspect the source again before adding another manual check. The purpose of the audit is a reliable record, not a longer reminder list.

Report the result with counts by requirement, source, and cause. Include the number of records repaired and the number still open. A narrative without counts cannot be compared with the next audit. The counts also show whether the repair reached the system or only the report.

Sources

References supplied with this guide.

Article source directory
SourcePublisherContext
U.S. Department of Labor, Occupational Safety and Health Administration Trainingosha.govPrimary reference supplied with this article.
U.S. Equal Employment Opportunity Commission, Recordkeeping Requirementseeoc.govPrimary reference supplied with this article.
U.S. Department of Labor, Recordkeepingdol.govPrimary reference supplied with this article.
National Institute of Standards and Technology, Cybersecurity Framework 2.0nist.govPrimary reference supplied with this article.