What AI features are included?
TL;DR: five features ship with every plan, all powered by Claude via AWS Bedrock with Titan embeddings for retrieval.
The five features
1. Ticket Summaries
Every new ticket gets a 2-3 sentence AI summary written into the ticket as a note. Agents see the summary in osTicket without leaving their normal workflow. Summaries also surface the customer’s primary issue and any urgency cues.
2. Recommendations
For each ticket, Flexivity searches your historical resolved tickets (semantic search via pgvector) and surfaces the 3 most relevant past resolutions. Recommendations appear as a panel inside the ticket — agents can copy a past response, adapt it, or use it as research.
3. Classification
Tickets are automatically classified into your existing osTicket help topics. If a customer’s ticket is filed under the wrong topic (or no topic), Flexivity proposes the right one. You can configure auto-apply or require agent approval.
4. KB Analysis
Periodically, Flexivity analyzes your ticket corpus and identifies patterns in resolved tickets that don’t have a corresponding KB article. The analysis surfaces “you should write a KB article about X” suggestions ranked by ticket volume.
5. KB Search
When a customer (or an agent) searches your knowledge base, Flexivity uses semantic search instead of keyword matching. Results match the intent of the query, not just the words. This is the same RAG pipeline that powers recommendations.
How tokens work
Every AI feature consumes tokens from your monthly allowance. The token cost per feature is documented on the billing page. You can disable individual features per instance if you want to control consumption.
Going deeper
- AI Summaries user guide — what they look like, how to tune them
- Recommendations user guide — the RAG pipeline explained
- Classification user guide — auto-apply vs. agent approval
- KB Analysis user guide — interpreting the suggestions
- Feature toggles — enable/disable features per instance