Permission-aware agents evidence gaps: what a review expects, what companies keep, and how to describe the shortfall

Short answer

The commonest gap is time: identity tools keep audit logs for days or weeks, so approval history disappears unless someone archived it. The others are loose links between approval and ticket, approver lists and role settings that changed without history, outcomes never recorded, and changes made under shared admin accounts. Each gap shrinks the number of assessable changes; none disqualifies a company alone. Describe gaps in an introduction as what is known, what is likely missing and who can confirm it.

Visual briefFull-size diagram
Six horizontal bars compare the evidence a review expects with what a typical company keeps: request tickets and action logs are mostly present, while approval history, approver lists, role settings versions and outcomes show dashed outlined segments where records are usually missing.
Illustrative pattern, not measured data. Bar lengths show a typical shape of what survives, not figures from any company.

This page is part of the guide to permission-aware agents and the second approver. That guide shows a complete trail for one account change. This page does the opposite job: it sets the evidence a review would expect against what companies usually keep, so an introducer can predict the shortfall and describe it without overselling.

The evidence a review expects

A reviewer judging whether an action had the right authority needs six kinds of evidence for each change. Missing any one weakens the case; missing the approval itself usually ends it.

EvidenceWhat it provesTypical system
Request recordWhy the change was needed, and who askedService desk or change ticket
Approval recordA different person granted authority, and whenIdentity tool, ERP approval workflow
Approver list historyWho was allowed to approve on that dateIdentity tool configuration, access reviews
Role or policy settings historyWhich rules applied: approval required, maximum duration, justificationIdentity tool configuration
Action logWhat was done with the authority, by whom, whenAudit log
Outcome recordWhether the change worked or was reversedTicket closure, follow-up ticket

What companies usually have and what is usually missing

The pattern below is a typical shape, not a measurement. It is what a reviewer should expect to find at a mid-sized company that never set out to keep approval evidence.

EvidenceUsually presentUsually missingWhy it goes missing
Request recordTickets for yearsHow the requester's identity was verifiedVerification is done on a call and not noted
Approval recordThe last few weeksEverything older than the tool's retention windowAudit logs expire unless archived
Approver list historyThe current listWho could approve on past datesLists are edited in place
Role or policy settings historyCurrent settingsEarlier settings and the date they changedConfiguration changes are not versioned
Action logRecent entries; shared-account entriesThe person behind a shared admin loginShared accounts hide the actor
Outcome recordTicket closedWhether the user's problem was solved, or a reversal linked to the changeClosure codes say "resolved" and nothing else

The approval row is the one to ask about first. Microsoft's documentation for Entra says audit logs are kept seven days on the Free tier and 30 days on P1 and P2, that they can be kept longer by routing them to Azure storage through Azure Monitor, and that expired data that was not archived cannot be recovered. A company that never set up archiving may have five years of tickets and one month of approvals.

How the gaps break a trail

Each gap breaks the authority story at a different point:

  • Missing context. The activation carries a reason and a ticket number, but the ticket was closed with no note of what was requested or how the requester was checked. A reviewer sees authority granted for an unknown purpose.
  • Loose linkage. Microsoft's PIM documentation says ticket information entered at activation is informational and not checked against any ticketing system. A mistyped number links an approval to the wrong request, or to nothing.
  • Unclear permissions. Approver lists changed and nobody kept the history, so it is impossible to say who was allowed to approve on a given day.
  • Moved versions. Settings such as whether approval was required or the maximum activation length were edited later, and the earlier values were not recorded. A change can look out of policy when it was within the rules of its day.
  • Shared accounts. A change made under a shared admin login has no individual actor, which defeats the purpose of a second approver.
  • Unrecorded outcome. The change was made, but nothing shows whether it solved the problem or was rolled back.

The blocked attempts matter too. Entra does not let approvers approve their own activation requests. Where a system keeps a record of a blocked self-approval, that record is the clearest evidence that the boundary held; ask whether such attempts are logged at all, because many summary reports never show them.

Describing a gap honestly in an introduction

An introduction that names the likely gaps builds trust and saves a review cycle. Keep it to three parts: what the contact confirmed, what is probably missing, and who can confirm the rest. Three Illustrative examples:

SituationHonest wording for the introduction
No archiving"The IT director confirms about 100 approval-gated changes a month and tickets back to 2019. Audit logs were not archived, so approvals older than 30 days are likely gone. The CIO can confirm."
Hand-typed links"Activations require a change number, typed by hand and not validated. Most should link; some will not. The service desk manager can estimate how often numbers are wrong."
Partial outsourcing"Approvals run in the company's own identity tool, but an MSP performs many changes under its own accounts. Who owns those action logs depends on the MSP contract."

Avoid three things: estimating what share of records survives, calling a gap "minor" before anyone has looked, and promising that a review will find a workaround. The reviewer decides how much is usable.

Where to look before calling a gap final

Some gaps close once the right person is asked. Audit logs may have been streamed to a security monitoring system even if nobody in IT remembers. Approver lists may survive in quarterly access review files. Outcome notes may sit in the ticket system's resolution field rather than the closure code. The historical ticket archive guide covers the service desk side of the same records, and the ERP replacement handoff checklist helps when approvals live in finance software due for retirement.

The company still needs 50+ full-time employees at peak (contractors excluded), documented operating history, rights to license and an authorized contact. Companies in the U.S., Canada and the EU qualify, as do companies in the UK, Australia, New Zealand, India, Singapore, South Africa, Latin America, the UAE and Saudi Arabia when at least 70% of the data is in English. Authority inside a task is a different test from the right to license; see de-identification versus permission to license. To score fit first, use the permission-aware agents qualification scorecard; the review packet contents list shows what the company would assemble later. More record families are in data types.

Next step

Refer a company. No cap on rewards. Default reward on a company's first qualifying transaction: the greater of $50,000 plus 5% and 15% of Actual Supplier Payout, with no cap; further qualifying transactions first agreed within 12 months of the first Closing pay 15%. Eligibility and payment conditions apply, including the USD 100,000 Minimum Transaction, and a signed Schedule A or addendum can differ. A lead, meeting or signature alone does not trigger payment.

If you can describe a client's approval records and their gaps in three honest sentences, register as a partner, check who qualifies, and read the program terms before you introduce the CIO or COO.

  1. Step 1Share your linkSend your personal link to a company you know.
  2. Step 2Company appliesThe company applies itself at /apply.
  3. Step 3Company signs and is paidThe company signs, delivers its data and is paid by SourceX under its agreement.
  4. Step 4You get your rewardthe greater of $50,000 plus 5% and 15% of the actual supplier payout, with no cap. A reward becomes payable after SourceX has paid the referred company under its signed agreement and the payout conditions in the program terms are met, including finance approval and your tax and payment details.

Common questions

Does a gap disqualify the company?

Rarely on its own. Each gap reduces how many approval-gated changes can be assessed. Only SourceX's review with the company can say how much survives, which is why an honest description of the gap helps rather than hurts.

Should I check the gaps myself before introducing?

No. Ask metadata questions, such as how long approvals are kept and whether logs were archived, and pass the answers on. Partners never open logs, tickets or exports.

Are these gaps specific to Microsoft Entra?

No. Entra is cited because its retention and ticket rules are publicly documented. ERP payment approvals, CRM discount approvals and banking dual-control tools show the same pattern of short retention and loose linkage.

Sources

  1. learn.microsoft.com/en-us/entra/identity/monitoring-health/refer…
  2. learn.microsoft.com/en-us/entra/id-governance/privileged-identit…
  3. learn.microsoft.com/en-us/entra/id-governance/privileged-identit…
  4. refer.sourcex.si/terms

Free resources

Know a company with valuable proprietary data?

Become a referral partner, get your link and introduce the company’s CEO (in an owner-run company, usually the owner). Referral partners can join from most countries. We can't work with partners in countries or regions under U.S. sanctions or on U.S. restricted-party lists. The companies you introduce should be based in the U.S., Canada and the EU, or in the UK, Australia, New Zealand, India, Singapore, South Africa, Latin America, the UAE and Saudi Arabia with at least 70% of their data in English.

Refer a company →

I own a business

Explore licensing your company's data directly to SourceX for AI training and evaluation. Start a short assessment; no uploads needed.

Start an assessment