Skip to main content

Troubleshooting

This page covers the most common issues when using AI/R Watch and how to resolve them. Most problems are related to authentication, organization access, data visibility, or the chat assistant.

No Data Shown for the Selected Period​

Symptom: The dashboard displays "No data found for the selected period" in one or more tabs.

Possible causes and solutions:

  1. The selected period has no recorded spend — try a different period (e.g., switch from "Current month" to "Previous month" or "3 months") to confirm whether data exists for other ranges.
  2. The organization has no AI spend data yet — if the organization was recently added to the platform, data may not have been ingested yet. Contact your platform administrator.
  3. The AI Only filter is active — in the Executive View, the AI Only toggle restricts data to AI-specific categories. Disable it to see all cloud spend.

Organization Selector Shows No Organizations​

Symptom: The organization selector in the header shows "No organization found" and the dashboard is empty.

Possible causes and solutions:

  1. Your account is not associated with any organization — request access from your platform administrator, who can add your account to the appropriate organization.
  2. Session issue — sign out and sign back in to refresh your session and reload the organization list.

Access Denied After Sign-In​

Symptom: After signing in, a message appears stating that your user does not have access to the application.

Solution: Your account was authenticated successfully, but it does not belong to any organization authorized to use AI/R Watch. Contact the responsible team to request access for your account.

LLM Tab Shows Partial or Empty Data​

Symptom: The LLMs tab displays a warning banner about partial data, or the Clouds and Models sub-tabs appear empty.

Cause: The detailed LLM view requires specific data access permissions. If those permissions are not configured for the current organization, model-level data will not be available.

Solution: Contact your platform administrator to verify that the data access configuration for your organization includes the required permissions for the LLM detail view.

Token Volume Section Not Visible​

Symptom: The Models sub-tab in the LLMs tab does not show the Token Volume section.

Cause: Token tracking requires an observability integration to be configured for your organization. If this integration is not set up, token data will not be available.

Solution: Contact your platform administrator to check whether the observability integration is configured for your organization.

AI Cockpit Finance Chat Returns an Error​

Symptom: The chat panel displays an error message such as "Failed to process your question. Please try again."

Possible causes and solutions:

  1. Temporary service issue — wait a few seconds and try again.
  2. Question is too complex — try rephrasing the question or breaking it into smaller, more specific parts.
  3. Request timed out — if the error mentions a timeout, the query took too long to complete. Try a simpler question or a shorter time period.

Chat Response Stops Mid-Answer​

Symptom: The assistant starts generating a response but stops before completing it.

Solution: Click the Stop button (■) if it is still visible, then try sending the question again. If the issue persists, start a new conversation using the New chat icon (✏️) and rephrase the question.

Theme or Language Preference Not Persisting​

Symptom: The dark/light theme or the selected language resets after closing and reopening the browser.

Cause: Theme and language preferences are stored in the browser's local storage. If local storage is cleared (e.g., by browser privacy settings or extensions), preferences will reset.

Solution: Re-apply your preferred theme and language from the account menu. If the issue recurs, check your browser's privacy or cookie settings to ensure local storage is not being cleared automatically.

Still Need Help?​

If the issue is not listed here or persists after following the steps above, contact your platform administrator or open a support request through the appropriate channel for your organization.