Troubleshooting AI results

Problem

Perslace Intelligence shows an incorrect or surprising plan state, currency, anomaly, recommendation, or simulation result, or it shows no finding when one was expected.

Likely causes

Stale or incomplete source context
Wrong scope, period, scheme, population, or baseline
A plan or assignment changed outside the reviewed context
A terminology mismatch between Perslace Intelligence, Copilot, or another AI label
A currency display defect
An Intelligence setting or access condition
An unavailable baseline or source record
A model limitation

Checks to perform

Record the Intelligence area, review time, scope, period, and displayed currency.
Compare the plan state with Commission Hub and effective employee assignments.
Compare material amounts with the organisation currency and an authoritative report.
Confirm that expected transaction data has completed its source refresh.
For an anomaly, check Revenue Contribution, Compensation Share, Rules Evaluated, and Rules Triggered where available.
For a simulation, check the scheme, objective, scope, period, baseline, population preview, and overlap warnings.
Confirm whether relevant Copilot or anomaly controls are enabled, without changing settings during diagnosis.
Restate the question with one clear scope and period, using no sensitive data.

Resolution steps

Do not act on the AI result.
Use authoritative scheme, assignment, transaction, report, and currency information for the current decision.
Preserve a sanitised record of the inconsistency.
If the issue was an unclear prompt, ask a narrower question and request scope, currency, assumptions, and limitations.
If the source context was stale, wait for the expected refresh and verify the source record before reviewing again.
Record the human conclusion and any follow-up owner.

Expected result

The AI result is either explained with verified source evidence or isolated as a reproducible inconsistency that has not affected an operational or financial decision.

Important notes

Human review required: AI output is not an authoritative financial record or approval.

Product review found No Active Plans in an AI recommendation while active plans were visible elsewhere.
Product review found INR in an AI surface while the organisation default currency was USD.
Do not save Intelligence settings merely because the page shows Unsaved changes; that state appeared without an intentional edit during review.

When to contact support

Contact Support when the inconsistency persists with aligned source context, the wrong currency or plan state remains visible, a valid baseline cannot be used, or the result could affect a financial decision.

Information to include in the support request

Intelligence area and time of review
Sanitised prompt or question
Scope, period, scheme context, and baseline where relevant
Configured and displayed currencies
Authoritative source comparison
Expected and displayed result
Redacted screenshot with confidential prompt content removed

Never include passwords, tokens, API keys, raw transactions, personal data, client-confidential information, or unrestricted reports.

Last updated: August 2026