Configuration documentation explains how to set it up. This is about finding out whether yours works.
In September 2026 I put the same ten questions to ChatGPT, Perplexity and Google AI Mode. Asked how to test effort certification in Workday, the first returned Workday's own configuration documentation, which explains how to configure it.
That is the closest thing that exists. The setup layer is well documented and the regulation layer is public, and almost nobody has written the layer in between.
The documentation standards are at 2 CFR 200.430(g). They were at paragraph (i) before the April 2024 revision, published at 89 FR 30136.
Paragraph (i) is now a different subject entirely. It covers Institutions of Higher Education, so a policy citing 200.430(i) for documentation standards is now pointing at the wrong text.
NIH cites 45 CFR 75.430(i), and that is still correct. HHS codifies the Uniform Guidance separately, so the same requirement carries two different citations depending on which book you are reading.
This is a five minute check and it is worth doing before an auditor does it for you. An institution whose effort policy cites a paragraph that no longer says what the policy claims has a documentation problem before anyone looks at a single certification.
Records charged to federal awards for personnel costs must meet all six. The wording below is the regulation's, condensed.
Supported by a system of internal control. The control has to give reasonable assurance that the charges are accurate, allowable and properly allocated.
Incorporated into the official records. A spreadsheet held by one administrator is not the official record of the institution.
Reasonably reflect total activity. Not exceeding one hundred percent of the activity the employee is compensated for.
Cover federally assisted and all other activity, on an integrated basis. Effort on a federal award cannot be documented in isolation from the rest of the person's work.
Comply with the institution's own accounting policies. Your own written policy becomes the standard you are tested against.
Support the distribution across activities where the employee works on more than one award or activity.
The regulation does not mandate a certification form or a particular effort reporting system, and NIH has said the same in its own guidance. Records may express activity as a percentage distribution.
Budget estimates alone do not qualify as support. They may be used for interim accounting where the estimating system produces reasonable approximations, significant changes are promptly recorded, and periodic after-the-fact reviews happen with adjustments made.
That last point is where a great many institutions actually sit. Charges are made on planned effort and the after-the-fact review is the control, which means the review is the thing to test rather than the certification screen.
None of these require access to anything confidential. All of them are answerable by the people who already run the process.
Build the list of people who should have certified from payroll actually charged to sponsored awards in the period. Then compare it to the list the system asked, because those two populations drift for reasons nobody notices until they are put side by side.
Record the certification date against the period it covers, not against the date the task was created. A certification completed eleven months after the period is a different control from one completed in the following month, and both show as complete.
Establish whether the person who clicked is the person the policy names, and whether delegation was used. The regulation requires a system of internal control, and a delegation nobody can explain is the point where that assurance stops.
Find out what was in front of the person at the moment they certified. If they saw a percentage with no underlying distribution, the attestation is about a number rather than about work performed.
Retroactive payroll adjustments are normal and they change the distribution somebody has already attested to. Establish whether anything in your configuration re-opens or re-triggers a certification when that happens, because in many tenants nothing does.
The certified distribution and the payroll posted to the award are two separate records and they are not automatically the same. A difference is not necessarily an error, but an unexplained difference is a finding.
An auditor asks for one certification from two years ago and everything behind it. Establish whether you can produce the certification, the distribution it covered, the basis for that distribution and the identity of the certifier, without anybody reconstructing it from memory.
This is the list of questions and the reasoning behind them. It does not give you the sampling approach, the scoring, the reconciliation logic or the evidence log those tests run on.
Those are the instruments and they are what the paid review is. The line is deliberate and it is the same line the free health check holds.
2 CFR 200.430, eCFR. Paragraph (g) carries the documentation standards. The page shows the amendment note.
NIH NOT-OD-18-108, standards for documentation of personnel expenses. States that no single method or form is required.
What evidence should support an effort certification is the companion to this page. It covers what belongs in the file behind each certification.
The free health check runs sixteen checks per area across four areas and costs nothing. The Sponsored Award Review is the scored version, forty-four checks with an evidence column.