Frequently asked questions and quick troubleshooting
Start with the symptom, then follow the linked procedure. The same missing button or empty result can come from the current context, permissions or plan.
Start with three checks
- Check the organization and whether you are working in real data or a scenario.
- Check the selected asset type, filters and archive state.
- Note the exact action and displayed error. For an access issue, identify the affected member and ask an administrator to use the permissions diagnosis.
For a support request, include useful steps and the error reference, without credentials or sensitive business data. See getting help.
Account and access
I can sign in, but cannot find my organization
A working account does not establish access to every organization. Check your invitation, active membership and the sign-in method accepted by the organization. An SSO requirement can make another method unsuitable. Follow organization access and sign-in.
I am an Administrator; why is a field hidden or read-only?
Administrative capabilities and business permissions are distinct. Check roles, access domains, the property and the selected context with permission diagnosis. Also check whether the property is calculated.
Does assigning an asset owner give them editing rights?
No. Ownership assigns responsibility for review. The member still needs the relevant reading or editing rights. An owner without edit access may use suggestions where available. See asset ownership.
Does sharing a folder give access to everything inside it?
Do not assume permanent inheritance from the folder. Folder or visualization access and underlying business data access are separate checks. Diagnose the specific resource and its data using content access.
Data and quality
An asset disappeared. Was it deleted?
First check organization, scenario, filters, type and permissions. Then inspect archived assets. A missing result or API 404 does not by itself establish permanent deletion. Search can help locate the asset in your accessible data.
My import created duplicates instead of updating records
Check the matching identifiers and mappings in the import preview. Names alone are not always unique or sufficient for matching. Stop further imports and inspect the result before retrying. Follow import troubleshooting, including the separate rules for relationships.
I changed an asset, but it is still stale
Editing and attesting are different actions. An owner must review and attest the asset. An explicit request for reattestation can also make it stale at any time, even when its type has no revalidation delay.
Why is a complete asset not considered fresh?
Completeness measures filled information; freshness concerns review. An asset can be 100% complete and still be stale or inaccurate. Conversely, some properties hidden from you by permissions can contribute to a score you cannot finish yourself. See completeness and quality indicators.
Why are some fields excluded from completeness?
Optional, calculated and metamodel-hidden properties do not count. Ownership, relationships and access domains do not count either. With no counted properties, the score is unevaluated, not 100%. See the completeness calculation.
Visualizations, changes and collaboration
Why do a chart, a view and an export show different totals?
Compare their type, context, filters, measure and relationship direction. Parallel relationships may contribute separately, while an asset filter can select an asset through one matching relationship. Check chart measures, filters and the chosen export scope before treating a difference as a calculation error.
Does copying a visualization copy its assets too?
A copied visualization normally reuses the same repository data. Presentation and underlying assets are different objects. Check move and copy and use a scenario when preparing supported data changes separately.
I edited a scenario. Why is the real repository unchanged?
Scenario changes remain in that context until you perform the applicable comparison and application workflow. Resynchronizing brings reference changes into the scenario; applying has a different direction. Follow compare and resynchronize and apply changes.
I sent a suggestion or received an AI proposal. Has it already changed the data?
A pending suggestion waits for an authorized reviewer to accept or reject it in Boldo’s suggestion workflow. An AI proposal is an action prepared in the chat, which you can apply or dismiss. Its action may change data directly or send a suggestion for a separate review. Check the recorded result through suggestions or AI proposal review.
A share link stopped working
Check expiration, deletion, organization sharing settings, the resource and the link's permissions. A temporary sharing switch is different from deleting the link. Use sharing administration; do not automatically publish a replacement with broader access.
I closed a cell editor. Why was the change not saved?
Cell edits can remain in a table draft after Enter, Escape or closing the editor. Saving a view configuration does not save its cell values. Use the table save/review workflow and verify the recorded result. See inventory editing.
How can I recover from a mistake?
Choose the operation that matches the change: local graphical Undo, visualization version restoration, asset restoration, or the supported scenario application undo. These are not interchangeable. Deleting an activity entry does not restore an earlier value.
Why did I receive no email or reminder?
Check the relevant notification preference, organization settings and whether there is eligible work. Manual reminders and scheduled summaries have different conditions. A sent confirmation does not prove inbox delivery. See review coordination.
Connections and AI
Can my external AI client change Boldo through MCP?
Yes, depending on its authorization. MCP can send suggestions for review in Boldo or make direct changes. Direct writes also require the organization setting, the member’s write level and write access chosen when connecting. These actions do not open AI proposal cards in Boldo’s chat. See MCP authorizations.
The API returned success. Why did some batch entries not change?
Lifecycle batches report applied and skipped; read each reason instead of relying only on HTTP 200. Other bulk operations have their own contracts. See API writes. After a timeout, read the state before repeating a write.
An endpoint test passes, but the assistant cannot use the model
An OpenAI-compatible test checks the models endpoint, not every generation capability. Check the selected model, service path, authentication and actual tool/image support. Configuration and a successful conversation are separate checks. See self-hosted AI.
A command is missing or disabled
Check your user type, business permissions, plan, organization activation and selected context. The access model distinguishes them; usage and limits explains capacity-related refusals. A shortcut does not bypass these conditions.