AI Ticket Intelligence
Every ticket is triaged the moment it arrives — categorised, prioritised, summarised and checked against what you have already solved — so the queue a technician opens is already sorted.
The problem
A ticket arrives as a paragraph of free text. Before anyone can work it, somebody has to read it, decide what it is about, decide how urgent it is, and check whether the same thing was solved last month. That triage tax is paid on every ticket, by the most expensive people in the business, and it is paid again every time the queue is re-read.
The duplicates are worse than the delay. Three users report the same outage as three tickets, three technicians pick them up, and the answer that resolved it in March is sitting in a closed ticket nobody thinks to search for.
What Solidlio does about it
Every ticket is analysed automatically at creation — no button, no queue to watch. The analysis produces a category, a priority with its reasoning, a summary, sentiment, business impact and suggested next steps, and is cached against the ticket so re-opening the panel costs nothing until the ticket actually changes. Alongside it, each ticket is embedded as a vector, so “tickets like this one” and “is this a duplicate of one already open” are similarity questions answered against your own history. Response drafting is a separate, deliberate action a technician takes, with audience and tone chosen per reply.
Capabilities
| Capability | What it does |
|---|---|
| Automatic triage | Runs on ticket creation in every portal, and as an automation-rule action |
| Full analysis | Category, priority with reasoning, summary, sentiment, urgency, business impact |
| Analysis caching | Re-reads are free; re-analysis happens only when the ticket itself changes |
| Duplicate detection | Flags likely duplicates while the ticket is still being typed |
| Similar tickets | Finds resolved tickets that match, with their resolutions |
| Knowledge suggestions | Surfaces relevant articles on the ticket, for staff and for the requester |
| Self-help deflection | Searches the knowledge base as an end user describes their problem |
| Response drafting | Generates a reply for a chosen audience and tone |
| Asset troubleshooting | Uses linked asset health, warranty and monitoring data for targeted next steps |
| Suggestion review | AI category and priority are proposed and applied on acceptance, not silently |
Built for MSPs and their clients
Staff triage tools follow the ticket, not the account boundary. An MSP working a managed client’s ticket gets the same analysis, response drafting and similar-ticket search it gets on its own — a managed client is a separate account, and the scoping recognises that rather than treating the client’s ticket as foreign.
| Organization | MSP | |
|---|---|---|
| Triage on their tickets | Automatic | Automatic, across every client they service |
| Response drafting | Their own staff | On any ticket they service |
| Knowledge suggestions | Their own + shared articles | Their own + the client’s, per the visibility model |
| Requester’s view | Deflection and article suggestions only | — |
How it works
- A ticket is created — through the MSP portal, the org portal, the customer portal or the API.
- Triage fires automatically. The analysis is charged to the right account, derived from the ticket’s organization.
- The result is stored on the ticket without marking it as edited, so “recently updated” queues are not disturbed by a background job.
- A technician opens the ticket. The copilot shows the cached analysis immediately; it re-runs only if the ticket changed since.
- Similar and duplicate tickets are offered from vector search over your own ticket history.
- The technician drafts a reply, choosing audience and tone, and edits before sending.
Credit cost
| Operation | Credits |
|---|---|
| Categorisation / triage / analysis | 1 |
| Response generation | 2 |
Limits
- Triage produces suggestions. Category and priority are written to AI fields on the ticket; applying them to the ticket’s own fields is an explicit acceptance.
- The AI category vocabulary and the ticket form’s category list are not the same set; accepting an AI category can produce a value the form does not offer.