Accreditation Evidence Agent
Maps what you already have to what the standard asks for
- 8hard rules
- Things it must never do, enforced in every run
- 5handover rules
- Conditions where it stops and gives the work to a person
- 10output fields
- A structured result, not a paragraph you have to re-read
- ≤8steps per run
- Bounded work: at most 6 tool calls, enforced by the runtime
Read-only by design. This agent has no path to sending, spending, publishing or committing anything. Where that limit is the product, removing it would remove the reason to trust the output.
Overview
Accreditation and inspection failures are rarely failures of substance. The teaching happened, the review happened, the improvement was made — and the institution cannot produce the document that shows it, or produces one that does not say what the standard asked.
So the year before a review becomes a scramble: staff reconstructing evidence for things they genuinely did, months of work producing nothing new for a single student. That scramble is the cost, and it repeats every cycle.
The difficulty is real. A standard is written in the accrediting body's language, evidence is produced in the institution's, and the mapping between them exists only in the head of whoever ran the last review. When that person leaves, it goes with them.
The failures are consistent. Evidence exists but nobody mapped it to the criterion it satisfies. Evidence is stale — a policy last reviewed outside the required cycle. A standard asks for a closed improvement loop and the institution can show the review but not what changed because of it. And criteria met by practice that was never written down, which is the most frustrating category because the work was done.
This agent produces the mapping and names each gap by type, because a missing document and an unmapped one need completely different responses. It writes no evidence and makes no compliance determination.
Benefits
The scramble replaced by a list
Of specific gaps, most of them small.
The mapping survives the person
Instead of leaving when they do.
Gaps found with time to fix them
Rather than in the month before submission.
What it does
Maps evidence you already hold
Most of the scramble is finding what exists, not creating it. The mapping lived in one person's head and left with them.
Names the gap type
Missing, stale, unmapped, or practice never recorded. Four different problems, and only one needs new work.
Checks the improvement loop closes
Standards ask what changed as a result. Institutions can show the review and not the change, and that is the classic finding.
Tests currency against the required cycle
A policy that exists but was last reviewed outside the cycle fails as surely as one that does not exist.
Finds practice nobody wrote down
The work was done. Only the record is missing, and that is a much smaller job than doing it again.
Writes no evidence
It never drafts a document to satisfy a criterion. Evidence generated to pass a review is exactly what a reviewer is looking for.
What this changes
What this changes about your week
- It takes over
- Mapping existing evidence to each accreditation criterion.
- Done today by
- A scramble in the year before review, reconstructing evidence for work already done.
- How often
- Every review cycle, from scratch.
- Typical manual effort
- Months of staff time producing nothing new for a single student.
Why Education in particular. The standard is written in the accrediting body's language and the evidence in the institution's. The mapping lives in one person's head, and leaves when they do.
Runs unattended
Started by you or by an event, and it finishes on its own. Nothing waits for someone to be at a desk.
The same standard every time
The two-hundredth item is held to the bar the first one was. Consistency is the part people cannot sustain.
It cannot act on its own
Accreditation Evidence Agent has no path to sending, spending or committing. That limit is why its output is safe to act on.
How it runs
This agent runs server-side through the PROMIVO runtime. Each run is logged step by step and every tool call is permission-checked before it executes.
- 1InputValidated against the agent input schema
- 2Modelclaude-opus-5 with automatic fallback
- 3Toolssearch_knowledge, extract_structured_data
- 4PermissionsRuns READ actions automatically; anything higher waits for approval
- 5OutputStructured result matching the output schema
- 6LoggingTokens, duration and every step recorded
Demo
Demo dataIllustrative sample output, abridged.
Input
{
"criteria": [],
"framework": "Institutional review 2026",
"reviewDate": "2027-05-10"
}Output
{
"mapped": [
{
"status": "partly-evidenced",
"criterion": "3.2 — The institution reviews assessment practice annually and acts on findings.",
"documents": [
"Assessment review 2025",
"Board minutes Nov 2025"
],
"partCovered": "The review happened and was reported.",
"passageCited": "The review reports the findings. The minutes note them as received.",
"partNotCovered": "Nothing evidences what changed as a result. The standard asks for the action, and this is the classic finding at every review."
},
{
"status": "practice-not-recorded",
"criterion": "5.1 — Student feedback informs curriculum design.",
"documents": [],
"partCovered": "",
"partNotCovered": "Staff describe doing this each term and it clearly happens. Nothing records it. Small to fix, and it must be written by the people who do it."
}
],
"escalate": true,
"disclaimer": "A mapping of the evidence retrieved against the criteria supplied. No evidence, policy or documentation has been written, no statement is made that any standard is met or that a review will be passed, and ambiguous criteria are reported rather than interpreted.",
"couldNotRetrieve": [
"Two committee minute sets referenced by the annual report are not in the evidence base."
],
"escalationReason": "Four criteria requiring a closed improvement loop can evidence only the review, and two gaps require a change in practice within 8 months of the review date.",
"ambiguousCriteria": [
"Criterion 4.4 asks for 'appropriate' external input without defining it. Whether the current arrangement qualifies depends entirely on that word."
],
"evidenceRetrieved": 214,
"closedLoopFindings": [
"Four criteria require review, action and change. All four evidence the review; one evidences the action; none evidence the change."
],
"needsADocumentOnly": [
"6 criteria are met by existing practice or existing evidence that was never mapped. Weeks of work, not terms."
],
"needsAChangeInPractice": [
"2 criteria require something the institution does not currently do. Those are measured in terms, and the review is in 8 months."
]
}Requirements
Integrations
No integrations required.
Before you start
- The standards or criteria you are assessed against
- Your existing evidence documents, in a knowledge base
- Review cycles and dates for each policy or process
- Requires the All-Access plan or higher
Use cases
Preparing for review
Twelve months out, when gaps are still cheap.
After a standard changes
What the new wording asks that the old one did not.
Continuous readiness
Checked each term rather than once a cycle.
A new quality lead
The mapping their predecessor never wrote down.
Pricing
All-Access
Billed monthly through your PROMIVO subscription. Cancel at any time.
Runs consume your plan allowance for agent executions and tokens. See plan limits
FAQ
Will it write our evidence?
No. Evidence generated to satisfy a criterion is precisely what a reviewer is trained to spot, and it is dishonest. It finds what exists, maps it, and names what genuinely has to be created by the people who did the work.
Does it say whether we will pass?
No. It makes no compliance determination. It reports which criteria have mapped, current evidence and which do not.
What if practice happens but is not documented?
That is a distinct finding, and usually good news. Recording something that already happens is far smaller than establishing it — but it must be recorded honestly, by the people who do it, not written by an agent.
Reviews
Reviews open once customers have run this agent.
Need this agent to work differently?
Tell us what to change and our team will scope a customised version for your business.
Customize this agent