The task
Service desk staff answer the same kinds of questions again and again, but each reply still has to fit the request in front of them. Writing it from scratch takes time, especially for complex requests or for customers who are already frustrated. Good answers often already exist somewhere: in knowledge base articles linked to the service space, or in the replies a colleague wrote when a similar request was closed. The page is written for agents who work through a queue of customer requests and want a first draft that already pulls in that material.
What they did
The page describes three ways to get a draft. In the first, the agent opens a request from the queue, chooses either an internal note or a reply to the customer and selects Suggest reply in the toolbar. Rovo writes a full reply from the linked knowledge base articles and from resolved requests that look similar, and it gives the articles priority over past requests. If neither offers useful context, it works from the current request and what the agent types. In the second, short suggested actions appear in the comment editor, such as a follow up, a request for more information or a proposal to close. They depend on the state of the request, for example how long the customer has been waiting, whether a service level target is close or whether the request was just reassigned. Picking one produces a reply that fits that action. In the third, the agent types a brief instruction, such as asking for error logs, and selects Enhance reply. Rovo expands it into a complete reply with the relevant details pulled in.
Where the approval sits
Rovo writes into the comment box but does not post anything to the customer. According to the page, the agent can insert, copy or edit each suggested reply before sending it, and that holds for all three ways of getting a draft. The agent also decides whether the text goes out as a reply to the customer or stays as an internal note. Because the draft leans on older resolved requests, checking that an earlier answer really fits the current case remains the agent's job.
Worth knowing
The source is Atlassian's own help page. It explains the feature and reports no results from any team. The quality of the drafts depends on how much context the service space holds, and Rovo may produce no draft at all when there are too few similar requests or linked articles. It may also refuse when a request appears to break Atlassian's usage rules. Drafting replies uses Rovo credits, so heavy use has a cost on top of the plan. The page does not say how similar requests are picked.
Source: support.atlassian.com · Sep 2026
Drafted with AI assistance and checked automatically. A person is responsible for the collection. How these entries are made
More in this category
- Inbox and scheduling
Copilot in Outlook turns an email thread into a meeting invite that waits for the sender
A Microsoft support page shows how Copilot in the new Outlook reads an email thread and prepares a meeting invitation with a title, an agenda, the attendees and the thread attached. The invitation stays open in the event form until the person has reviewed it, changed what is needed and pressed Send.
Read the entry - Inbox and scheduling
A rules file guides Claude Cowork to draft email replies each morning
A professional writes a rules file that tells Claude Cowork how to sort the morning inbox and draft replies, then reads a one-page summary. The Gmail connector can only create drafts, so the author reads and sends every reply.
Read the entry - Inbox and scheduling
A web agency sorts its Outlook inbox with AI and approves the drafts
A small web-design agency used an n8n workflow to classify the messages in its Outlook inbox and draft replies for the harder ones. The owner approves each of those drafts with one click, and the workflow logs its actions to a Google Sheet.
Read the entry
Ninety minutes live on your tools. One automation running before we hang up.