Skip to Content
IntegrationsosTicketFeature Configuration

Feature Configuration

Once osTicket is connected and active, you can enable and configure individual AI features from the Flexivity AI administration console. Each feature can be turned on or off independently per instance.

Accessing Feature Settings

  1. Log in to the Flexivity AI administration console .
  2. Navigate to Instances and select your osTicket instance.
  3. Scroll down to the Features section.

Each feature has its own toggle and configuration options.

AI Summaries

AI Summaries automatically generate a concise summary of each ticket when it is created or updated. Summaries appear in the osTicket ticket view, giving agents a quick overview without reading the full thread.

How It Works

  1. A ticket event (create or update) fires in osTicket.
  2. The plugin captures the event and queues it.
  3. Flexivity AI picks up the event, scrubs PII, and sends it to the cloud.
  4. The AI pipeline generates a summary using the ticket subject, description, and thread messages.
  5. The summary is sent back and written to osTicket via the plugin API.

Enabling AI Summaries

  1. In the Features section, find Ticket Summary.
  2. Toggle the feature to Enabled.

New ticket events will begin generating summaries within a few seconds. Existing tickets are not retroactively summarized (use Historical Sync for that).

Note: AI Summaries are generated on ticket create and update events. If a ticket has many rapid updates, only the most recent event is processed to avoid redundant summaries.

Ticket Recommendations

Ticket Recommendations use retrieval-augmented generation (RAG) to search your historical ticket data and suggest similar resolved tickets that may help agents resolve the current issue.

How It Works

  1. When a ticket event is processed, the system generates an embedding of the ticket content.
  2. The embedding is compared against your historical ticket embeddings using vector similarity search.
  3. The most relevant resolved tickets are retrieved and used to generate a recommended resolution.
  4. The recommendation is written back to the ticket view in osTicket.

Enabling Ticket Recommendations

  1. In the Features section, find Ticket Recommendation.
  2. Toggle the feature to Enabled.

Note: Recommendations are most useful after a Historical Sync has been completed, as the system needs a corpus of resolved tickets to draw from. With a small or empty knowledge base, recommendations may be limited.

Historical Ticket Sync

Historical Sync imports your existing ticket archive into the Flexivity AI knowledge base. This builds the vector embeddings that power Ticket Recommendations, so the system has full context from your ticket history rather than starting from zero.

Running a Sync

  1. In the Features section, find Historical Sync.
  2. Click Start Sync.
  3. The sync job processes tickets in batches. Progress is displayed in the administration console showing:
    • Total tickets to process
    • Tickets completed
    • Current batch progress
    • Estimated time remaining

Sync Behavior

  • The sync processes all resolved and closed tickets in your osTicket database.
  • Each ticket’s subject, description, and thread messages are embedded.
  • PII is scrubbed before any data leaves your environment.
  • The sync is idempotent — running it again skips tickets that have already been processed.
  • Large archives (tens of thousands of tickets) may take several hours. The sync runs in the background and does not affect osTicket performance.

Warning: For very large ticket archives (50,000+ tickets), the initial sync may take significant time. Plan to run the first sync during a low-traffic period. Subsequent syncs are incremental and much faster.

Monitoring Sync Status

The administration console shows the current sync job status:

  • Running — The sync is actively processing tickets.
  • Completed — All tickets have been processed.
  • Failed — The sync encountered an error. Check the job details for the failure reason.

If a sync fails partway through, restart it. It will resume from where it left off.

AI Ticket Classification

AI Ticket Classification automatically categorizes incoming tickets by comparing them against your historical ticket data. It assigns tickets to the appropriate help topic, reducing manual triage work for agents.

How It Works

  1. A new ticket event fires in osTicket.
  2. The system generates an embedding of the ticket content and compares it against historical tickets with known help topic assignments.
  3. If the confidence score exceeds the auto-assign threshold, the help topic is applied automatically.
  4. If the confidence score is low, classification takes a second path where the LLM estimates the appropriate category based on the help topic list and the internal descriptions taxonomy, which it compares against the intent of the ticket. Agents can then accept a suggestion from the LLM with a single click or override it manually.
  5. Accuracy is tracked at ticket closure — if the agent changes the help topic, the system records the correction.

Enabling AI Classification

  1. In the Features section, find Ticket Classification.
  2. Toggle the feature to Enabled.

Configuration Options

  • Auto-assign threshold — The minimum confidence score (0.0 to 1.0) required for the system to automatically assign a help topic. The default is 0.85. Lower values auto-assign more aggressively; higher values fall back to suggestions more often.
  • Sync Help Topics — Seeds the cache of classification categories from your helpdesk’s help topics. This is also updated automatically when first connected and once daily.

Note: Classification requires a meaningful corpus of historical tickets with help topic assignments. Run a Historical Sync before enabling classification for the best results.

Knowledge Base Gap Analysis & Authoring

Knowledge Base Gap Analysis examines your resolved ticket history to identify recurring topics not covered by existing FAQ articles. It then generates draft articles that administrators can review, edit, and publish to osTicket.

How It Works

  1. Resolved tickets are clustered by topic similarity to find common issue patterns.
  2. Clusters are compared against published FAQ article embeddings to identify coverage gaps.
  3. For each gap, the system generates a draft KB article using AI, drawing from the most relevant resolved tickets.
  4. Administrators review and publish drafts from the administration console.

Enabling KB Gap Analysis

  1. In the Features section, find KB Gap Analysis.
  2. Toggle the feature to Enabled.

Configuration Options

  • Minimum cluster size — The minimum number of tickets a topic cluster must contain to be flagged as a gap. The default is 5. Higher values surface only the most common gaps.
  • Coverage threshold — The similarity score above which a cluster is considered “covered” by an existing FAQ article. The default is 0.80. Lower values are more strict about coverage.
  • Article format — Choose between FAQ (short Q&A format) or KB Article (longer form knowledge base article).

Note: KB Gap Analysis works best after a Historical Sync, as it needs a substantial corpus of resolved tickets to identify meaningful patterns. Published FAQ articles must also be embedded (see KB Article Embedding in the Features section) for accurate coverage detection.

Knowledge Base AI Search adds a floating search widget to your public-facing osTicket pages. End users can type natural language questions and receive streaming AI-generated answers grounded in your published FAQ articles.

How It Works

  1. The search widget is injected into public osTicket pages via a JavaScript snippet included by the plugin.
  2. When an end user submits a question, it is sent to the Flexivity AI search API.
  3. The API performs a vector similarity search against your published FAQ article embeddings.
  4. Relevant articles are used as context to generate a streaming AI answer, with source links back to the original FAQ articles.
  1. In the Features section, find KB AI Search.
  2. Toggle the feature to Enabled.

Configuration Options

  • Response Style — Choose between Standard, Concise, or Friendly Customer Service to control the tone and length of AI-generated answers.
  • Temperature — Controls the creativity of the AI response. Lower values produce more focused, deterministic answers; higher values produce more varied responses.
  • Similarity Threshold — The minimum similarity score required for an FAQ article to be considered relevant to the user’s question. Higher values return fewer but more precisely matched results.

Note: KB AI Search requires published FAQ articles to be embedded. If no articles are embedded, the widget will return no results. Enable KB Article Embedding and ensure your FAQ articles are published before activating this feature.

AI Prompt Settings

For advanced users, the AI prompt templates used for summaries and recommendations can be customized per instance. This allows you to tailor the AI output to your organization’s tone, terminology, or specific instructions.

Accessing Prompt Settings

  1. In the administration console, navigate to your instance.
  2. Scroll down to the Features section.
  3. Select the feature whose prompt you want to edit: Summary, Recommendation, Classification, or KB Article Generation.

Editing a Prompt

Each feature has a default prompt template that works well for most use cases. To customize:

  1. Click Edit on the prompt you want to modify.
  2. Modify the prompt text. The editor shows available template variables (e.g., {{ticket_subject}}, {{ticket_body}}, {{thread_messages}}).
  3. Click Save Draft to save without activating.
  4. Click Publish to make the new prompt active.

Note: Prompt changes take effect immediately for new events. In-flight events already queued for processing will use the prompt version that was active when they were enqueued.

Reverting a Prompt

Each prompt edit creates a new version. You can view the version history and revert to any previous version:

  1. In the prompt editor, click Version History.
  2. Select the version you want to restore.
  3. Click Revert to This Version and then Publish.

Feature Health Monitoring

The administration console provides operational metrics for each instance on the Dashboard tab:

  • Jobs Today — Number of AI processing jobs executed in the current day.
  • Failure Count — Number of failed jobs. A non-zero count warrants investigation.
  • Agent Status — Current connection state (Connected, Disconnected, Stale Heartbeat).
  • Last Event Processed — Timestamp of the most recent successfully processed ticket event.

If the failure count is elevated, check the instance’s job history for error details. Common causes include transient cloud errors (which are retried automatically) and persistent issues like an unreachable plugin endpoint.

Next Steps

With features configured, your integration is fully operational. For ongoing operations:

Last updated on
Connecting osTicketIntegration Approach