Skip to main content

Smart Response

Drafts reply messages the agent can send to the customer, grouped by intent. It's a context-aware template generator - the AI reads the ticket and writes drafts for different scenarios (acknowledge, ask for info, propose a fix, close out). The agent picks the right intent, edits if needed, and adds it as a public note.

Smart Response panel with three accordion cards stacked vertically, each labelled with an intent ("Acknowledge", "Ask for more info", "Propose fix"). One card is expanded showing a multi-paragraph draft response with Edit and Add as Note buttons

What the agent sees​

A list of accordion cards. Each card is one intent the AI judged plausible for this ticket. Expanding a card reveals:

  • The drafted response in full, written in the agent's voice.
  • An Edit button - opens the text in an editable box.
  • An Add as a Note button - appends the response as a public ticket note in one click.

The intents are generated dynamically per ticket. A new-laptop-request ticket might get Acknowledge, Confirm specs, Send pickup instructions. An angry "still broken" ticket might get Apologise, Escalate to L2, Propose workaround.

What it reads​

  • Ticket subject, description, notes
  • Ingested knowledge (so SLA language, policy boilerplate, FAQ answers come out right)
  • Past tickets (so the wording matches the tenant's usual style)

When it refreshes​

On every public note - whether the agent or the user wrote it. The drafts adapt as the conversation evolves. A response that made sense before the customer's last reply may no longer fit; the panel regenerates so the agent always sees current drafts.

What gets added to the ticket​

Only what the agent clicks. Smart Response writes drafts to the panel, not to the ticket. Until the agent presses Add as a Note (after editing or not), the customer never sees it.