Skip to content
ticketagent

Support ticketing, from first message to resolved

Every ticket in Ticketagent follows the same seven steps. Here is what happens at each one, what your team configures, and where the AI agent helps.

Try the AI agent
  1. 1.

    A message arrives and becomes a ticket

    A customer writes to your support address, opens the chat widget, fills in a web form, or your product creates a ticket through the API. Ticketagent turns it into a ticket in the right inbox and attaches it to that customer's history, so the agent sees earlier conversations at a glance.

    • Email arrives through a forwarding address you set once
    • Chat and form widgets drop a snippet on your site
    • Replies to an existing thread join the same ticket instead of opening a new one
  2. 2.

    The AI classifies it

    Before any rule runs, the AI reads the message and sets four fields: category, priority, sentiment and language. It also writes a one-line customer intent, such as "refund for a duplicate charge". Your team sees the fields and can correct them, and the correction is kept in the ticket history.

  3. 3.

    Rules route it to a team and an agent

    Routing rules use those fields together with the channel, the customer, tags and business hours. A rule reads like a sentence: when intent is refund and plan is Team, route to Billing with priority high. Inside the team, round-robin or load-balanced assignment picks the agent.

    WhenThen
    Category is billing Route to Billing
    Customer tag is vip Set priority to urgent
    Channel is chat outside business hours Send to the next shift
    Sentiment is negative and priority is high Notify the team lead

    Every routed ticket shows the rule that fired, so nobody wonders why it landed where it did. More in help desk automation.

  4. 4.

    The SLA clock starts

    The ticket gets the SLA policy that matches its priority and the customer's tier. The first response clock and the resolution clock run on business hours, pause while the ticket waits on the customer, and show in the queue as a countdown. When a clock gets close to zero the ticket moves to the Breaching view and the assignee gets a warning. If it runs out, the ticket escalates to a lead.

  5. 5.

    The AI agent drafts or resolves

    The AI agent searches your knowledge base and your approved replies, writes a draft, and shows the article it used. Then it makes one of two decisions.

    • Auto-resolve the intent is on your approved list and the confidence is above your threshold. The agent sends the reply and marks the ticket Auto-resolved.
    • Hand to agent anything else. A person gets the ticket with the draft and a two-line summary of what the customer needs.

    You decide which intents are approved and how confident the agent must be. Read about the AI customer service agent.

  6. 6.

    A person handles the rest

    Agents work the queue with the draft already written. They check the cited source, edit, add an internal note or @mention a colleague, and send. If the ticket is a bug, they create a Jira or GitHub issue from the ticket and the issue status shows on it.

  7. 7.

    Everything is measured

    When a ticket is solved, the customer gets a short CSAT survey. Reports show volume, first response time, resolution time, SLA attainment, AI resolution rate and CSAT, by team, channel and tag. Resolved tickets also feed the knowledge base: Ticketagent suggests new articles from answers that keep repeating.

What your team sets up once

  • Inboxes and the support addresses that forward into them
  • Teams, business hours and holidays
  • SLA policies per priority and customer tier
  • Routing rules and assignment method
  • Knowledge base articles the AI agent may cite
  • The intents the AI agent may resolve alone, and its confidence threshold

Most teams start with one inbox, one SLA policy and a handful of articles, then add rules as patterns show up in the queue. The ticketing system page covers statuses and views, and the knowledge base software page covers articles.

Support ticketing questions

A small team connects its support address, writes a few articles and turns on AI drafts in an afternoon. Rules and SLA policies grow from there.

It sets priority when the ticket arrives. Agents can change it, and your rules can override it.

A normal reply from your support address. Auto-resolved tickets can include a line that invites the customer to reply if they still need help, and a reply reopens the ticket.

Yes. With auto-resolve off, the AI agent still drafts every reply and a person sends it.

See support ticketing on a real message

Run the AI agent on one of your own customer messages and watch it classify, route and draft.

Try the AI agent

Help desk with an AI agent, from $29 per agent.