Troubleshooting and FAQ
Recommended troubleshooting approach
The most effective troubleshooting model in AssureGrid is not to guess whether the product is broken; it is to verify the few factors most likely to explain the behavior. In most cases, users should start by checking visible page state, processing status, save state, active version, selection context, and permissions.
Common troubleshooting situations
| Symptom | What to check first | Likely next action |
|---|---|---|
| A page looks outdated or incomplete | Refresh the page, confirm the correct workspace or audit context, and verify that the most recent edits were saved. | Reload the view, reopen the module, or compare against the expected saved version. |
| Processing seems stuck | Check whether the action is still running in the background, whether a queue or status indicator is present, and whether the input is unusually large. | Wait briefly, refresh status, then retry only if the prior attempt clearly did not complete. |
| A download does not match expectations | Confirm the active version, selected rows or filters, and whether the displayed content is the content intended for export. | Recheck selections, save recent changes, and download again from the correct stage or version. |
| AI output is different from what was expected | Review the selected context, prompt wording, and whether the result is a suggestion versus an applied change. | Refine the request, edit manually if needed, and save only after approving the final version. |
| Evidence or attachment is missing | Verify permissions, folder context, selection scope, and whether the evidence item was uploaded or linked successfully. | Retry from the correct folder or control area and escalate if authorized users still cannot access the file. |
| A user cannot complete an action | Check role-based permissions, required fields, unresolved validation messages, and whether the stage has prerequisite steps. | Complete missing fields, confirm permissions, then retry or contact support if access should already exist. |
Module-specific troubleshooting guide
Different parts of AssureGrid surface different issue patterns. The most effective troubleshooting starts by identifying the module, then checking the few conditions most likely to explain the behavior before retrying or escalating.
| Module | Typical issue pattern | Recommended user check |
|---|---|---|
| Support Center | Submission appears not to go through, or users are unsure whether the request was captured. | Look for the confirmation state before submitting again. |
| Control Inventory | Users may not see the expected AI or table result because the wrong rows, columns, or cells were selected, or no save was completed after review. | Confirm selection scope and save state. |
| Audit Planning | Generated planning outputs may appear inconsistent when users compare a draft step against another stage or expect a final approval behavior. | Check which planning step is active and whether the visible draft is the current one. |
| Data Refinements | Users may confuse an AI suggestion with an applied refinement or may expect an unsaved comparison state to persist. | Confirm whether the original, accepted, or manually edited version is actually saved. |
| Evidence Management | Bulk evidence downloads or individual file access can fail due to selection or folder-context mistakes. | Recheck the selected evidence items and control folder context. |
| Audit Execution | Workpaper exports may be mistaken for final reporting output or may reflect a different template-driven structure than expected. | Confirm that the user wants the Excel workpaper, not the final report. |
| Issue Management | Issue wording, severity-related edits, or offline issue exports may not reflect unsaved changes. | Review the issue row state and save the intended version before downloading. |
| Audit Report Generation | The PDF may not match expectation if a different version is active, restored, or currently visible on screen. | Verify the active report version before downloading the final PDF. |
When support should be contacted Users should escalate when a workflow remains blocked after basic checks, when the same issue is reproducible, when multiple users are affected, or when the problem appears related to permissions, missing data, or unexpected system behavior that the user cannot resolve independently.
Frequently asked questions
The FAQ section is intended for quick-answer scenarios. These are not deep diagnostic playbooks; they are concise answers to the questions users ask most often when moving through the platform.
How do I know whether a change was saved?
Users should rely on the visible save behavior in the module they are working in. A suggestion, preview, or AI response is not automatically the same thing as a saved record.
Why does a download look different from what I expected?
The most common reasons are version mismatch, unsaved edits, incorrect filters or selections, or downloading from the wrong module stage.
Does processing always happen instantly?
Not always. Some workflows may use background processing, especially when generating, refining, exporting, or assembling larger outputs. Users should confirm status before retrying.
Can AI suggestions be treated as final automatically?
No. AI assistance supports drafting, explanation, and refinement, but user review and approval still determine what becomes part of the audit record.
What should I do before reporting a bug?
Confirm the workspace context, save state, visible version, permissions, and whether the issue can be reproduced consistently. Then submit the clearest possible description to support.
Why can I see something but not download or update it?
This often points to permissions, role limitations, or stage-specific controls. Visibility does not always imply edit or export rights.
When should I escalate instead of retrying?
Escalate when the same issue persists after basic checks, when multiple users are affected, when a key audit deadline is blocked, or when the problem appears to involve permissions or missing data that users cannot correct themselves.