Capabilities

Flexivity AI provides a set of integrated capabilities that work together to reduce resolution time, improve consistency, provide better customer service, and give administrators visibility into AI-assisted helpdesk operations.

AI Ticket Summaries

Every ticket conversation is automatically summarized as it progresses. When an agent opens a ticket, they see a concise, up-to-date summary of the entire conversation thread rather than reading through dozens of individual messages.

How it works:

The helpdesk app extension captures ticket events (new messages, internal notes, status changes), scrubs PII, and sends them through the processing pipeline
The AI uses a customizable prompt to generate a structured summary that covers the reported issue, steps taken, and current status
The summary is written back into your helpdesk app and displayed alongside the ticket conversation, including customer sentiment for additional context
Summaries update automatically as new messages are added to the ticket

Use case: An agent picks up a ticket that has been passed between three team members over two weeks. Instead of reading 25 messages, they read a three-paragraph summary and immediately understand the current state.

Smart Recommendations

When a ticket is processed, Flexivity AI searches your historical ticket database for similar past issues and surfaces relevant solutions. This is powered by retrieval-augmented generation (RAG) using vector embeddings of your own ticket data.

How it works:

The helpdesk app extension picks up new ticket creation events and delivers them through the processing pipeline
Ticket content is converted and stored in a vector database for fast similarity search
When a new ticket arrives, the system finds the most similar historical tickets and uses them as context for generating recommendations
Recommendations reference specific past tickets sorted by relevance, so agents can verify the suggested approach

Use case: A user reports a VPN connectivity issue. Flexivity AI finds three similar tickets from the past six months, one of which documents a configuration change that resolved the same symptoms. The agent applies the fix in minutes instead of troubleshooting from scratch.

Historical Ticket Sync

To build an effective knowledge base from day one, Flexivity AI can import and embed your existing ticket history. This ensures that recommendations are available immediately, not just after months of accumulating new tickets.

How it works:

An administrator initiates a sync job from the admin console, selecting a date range
The agent reads historical tickets from the helpdesk, scrubs PII, and sends them to the cloud for embedding
Tickets are processed in batches to avoid impacting helpdesk performance
Progress is tracked in the admin console, with visibility into sync status and any errors

Use case: An organization with 50,000 historical tickets activates Flexivity AI. Within hours, their entire ticket history is embedded and available for similarity search, giving agents immediate access to institutional knowledge.

Feedback Analytics

Helpdesk agents can rate AI-generated summaries and recommendations directly within the helpdesk interface. This feedback is aggregated and presented in the admin console, giving administrators data-driven insight into AI quality.

How it works:

Each AI-generated summary and recommendation includes a feedback mechanism (thumbs up/down)
Feedback is completely anonymized — no user information is captured or tracked
The admin console displays trends over time: accuracy rates and per-feature breakdowns

Use case: An administrator notices that quality scores are low for summaries. After investigation, they determine that agents are saying summaries are sometimes too long. They update the prompt to shorten them, and following this change the quality scores improve.

AI Ticket Classification

When a new ticket arrives, Flexivity AI can automatically categorize it based on similarity to your historical ticket data. This reduces manual triage and ensures tickets reach the right team faster.

How it works:

The system generates an embedding of the incoming ticket and compares it against historical tickets with known help topic assignments
Two operating modes: auto-assign applies the help topic automatically when confidence exceeds a configurable threshold, and suggest displays ranked help topic suggestions for borderline cases
Agents see classification suggestions directly on the ticket detail page and can apply them with a single click
Classification accuracy is tracked at ticket closure — if the agent changes the help topic, the system records it for quality monitoring

Use case: A new ticket about printer issues is automatically assigned to the “Hardware” help topic, saving the agent from manually categorizing it.

Knowledge Base Gap Analysis & Authoring

Flexivity AI analyzes your resolved ticket history to identify recurring topics that are not covered by existing knowledge base articles, then generates draft articles to fill those gaps.

How it works:

Resolved tickets are clustered by topic similarity to find common issue patterns
The system syncs all of your KB articles at initial setup, and then synchronizes changes to published articles to maintain a current record of your KB
Clusters are compared against published FAQ article embeddings to detect coverage gaps — topics with many tickets but no matching FAQ are flagged
For each gap, the system generates a draft KB article using AI, drawing content from the most relevant resolved tickets in that cluster. There are two modes for articles, either FAQ Q&A format or longer form KB articles
Administrators generate and publish drafts directly to osTicket as FAQ articles from the admin console, where they can then be further edited and published

Use case: After analyzing six months of tickets, the system identifies that “VPN setup” is a frequent topic with no KB article. It generates a draft article from the 15 most relevant resolved tickets.

Knowledge Base AI Search

A floating search widget on your public-facing osTicket pages lets end users ask natural language questions and receive AI-generated answers grounded in your published FAQ articles.

How it works:

The widget appears on public osTicket pages (knowledge base, home page, and ticket submission form)
End users type a question in natural language and receive a streaming AI-generated answer
Answers are synthesized from your published FAQ articles, with source links so users can read the full article
By helping users self-serve, the widget can deflect tickets before they are created

Use case: An employee searching for “how to connect to VPN from home” gets a comprehensive answer synthesized from the company’s FAQ, with a link to the full article.

Deployment Modes

Flexivity AI supports two deployment models to match different infrastructure and compliance requirements.

Self-Hosted Agent (On-Premise)

The standard deployment model. A Docker-based agent runs in your environment alongside your helpdesk application. The agent handles PII scrubbing locally before sending any data to the cloud.

Characteristics:

PII never leaves your network — scrubbing happens on-premise before transmission
Outbound WebSocket connection only over TLS on port 443 — no inbound firewall rules required
Runs as a lightweight Docker container (Docker Engine 20.10+, Docker Compose v2)
Automatic updates via container image pulls

Direct Connection (Cloud-Managed)

For organizations whose helpdesk application is internet-accessible, Flexivity AI can connect to it directly from the cloud. In this mode, PII scrubbing occurs at the initial ingestion point before any downstream processing.

Characteristics:

No on-premise infrastructure required beyond the helpdesk application itself
The cloud service connects directly to the helpdesk app for new events
PII scrubbing occurs at the initial ingestion point before data enters the processing pipeline
Suitable for cloud-hosted or SaaS helpdesk deployments where running a Docker agent is not practical