AI Report Reader automatically analyzes inspection report PDFs and images uploaded to the Inspections module and generates draft maintenance tickets for users to review and approve. When a file is uploaded as an external protocol to an inspection, the AI detects potential defects and surfaces suggestions in the inspection view - each pre-filled with a ticket title and description. The user reviews, edits if needed, and confirms creation.
📌 Note: Every ticket requires explicit user confirmation. AI does not create tickets automatically. Attachments added directly to a Note are not analyzed - only files uploaded as an external protocol to an inspection are processed. |
How it works
When a PDF or image file is uploaded as an external protocol to an inspection, AI Report Reader scans the document for potential defects. Analysis is triggered automatically, immediately after the file is uploaded - there is no manual trigger. Suggestions appear in the inspection view, each pre-filled with a ticket title and description derived from the document content.
When AI Report Reader has detected issues and generated suggestions for an inspection, a dismissible awareness banner appears at the top of the inspection detail page, prompting users to review the AI findings. Clicking "Show more" navigates directly to the relevant note, switching to the correct tab automatically if needed. Once dismissed, the banner does not reappear for that user.
The user can review and edit each suggestion before confirming ticket creation. Every AI-generated ticket is automatically linked to the source inspection protocol for full traceability.
Each suggestion can be individually accepted or ignored. Ignoring a suggestion has no effect on the inspection or the report. Once a ticket is created from a suggestion, the suggestion card shows the ticket number instead of the create button, preventing duplicate creation.
📌 Language of suggestions: Suggestions are generated in the UI language of the user who uploaded the file - not the language of the document. For example, if a Polish-language user uploads a German report, suggestions will appear in Polish. If a different user reviews the suggestions later, they will see them in the uploader's language. The content is fully editable before ticket creation. |
Key benefits
Reduces manual review time significantly.
Fewer missed defects - AI scans the entire document or image.
Faster defect-to-repair cycle: Report → suggestion → ticket in one flow.
Full human control: AI suggests, user approves.
Every AI-generated ticket is automatically linked to the source inspection protocol for full traceability.
AI suggestions on the inspection list
The inspection list also surfaces AI Report Reader activity so suggestions are not missed by only opening individual inspections:
Each row with AI-generated linked ticket suggestions shows a dedicated icon; hovering over it displays a tooltip confirming suggestions are available for that inspection.
A dismissible banner at the top of the inspection list shows the total number of AI-generated ticket suggestions across all inspections on the instance, alerting coordinators and FM managers with linked-ticket-creation permissions that suggestions are waiting for review.
No additional configuration is needed beyond having AI Report Reader activated.
Rating suggestion quality
Users can rate each AI-generated ticket suggestion with a thumbs up or thumbs down directly in the inspection view. A negative rating opens a follow-up form with predefined categories to explain what went wrong, and all ratings are captured automatically as feedback for the product team. Available to all users who can view AI suggestions; no configuration is required.
Permissions
Section | Name | Description |
Admin | ApplicationSettings | Required for administrators to access and manage AI settings in Application Settings → AI Functions. |
End users do not need any special AI-related permissions. Standard Helpdesk ticket creation permission is sufficient to use the feature.
Users without admin access can request activation directly from the feature - the admin receives a notification with a direct enable link.
Limitations
AI Report Reader only analyzes files uploaded after the feature was enabled on the instance. There is no retroactive analysis of previously uploaded files.
Each instance has a configurable daily processing limit (Daily limit for AI-assisted ticket auto-generation in Application Settings → AI Functions). Files uploaded beyond the limit are queued and processed the following day.
📌 Note: If analysis fails - due to a technical error, timeout, or the daily processing limit being reached - the inspection view will show a warning indicator. There is no manual retry; the system will automatically attempt to re-analyze the file the following day. |
If an uploaded report is too large to be analyzed, the report itself now shows a clear message explaining that it cannot be analyzed and inviting the user to upload a smaller file - instead of the file being silently retried every day with no explanation. This applies to all users of AI Report Reader and requires no configuration.
Remember that!
📌 Attachments added directly to a Note within an inspection are not analyzed. Only files uploaded as an external protocol to an inspection are processed by AI Report Reader. |
📌 Even when no issues are detected, the system always displays a disclaimer recommending manual review of the inspection report. |
📌 AI is a productivity aid, not a replacement for human review. It is not guaranteed to detect 100% of defects. |
