Showing results for 
Search instead for 
Did you mean: 
Sign up Log in

Earn badges and make progress

You're on your way to the next level! Join the Kudos program to earn points and save your progress.

Deleted user Avatar
Deleted user

Level 1: Seed

25 / 150 points

Next: Root


1 badge earned


Participate in fun challenges

Challenges come and go, but your rewards stay with you. Do more to earn more!


Gift kudos to your peers

What goes around comes around! Share the love by gifting kudos to your peers.


Rise up in the ranks

Keep earning points to reach the top of the leaderboard. It resets every quarter so you always have a chance!


🤖 Post-AMA Recap: Virtual agent and Atlassian Intelligence in Jira Service Management

👋 Hello Atlassian Community!

Blake here from the Jira Service Management team. In case you missed it, we recently hosted an Ask Me Anything (AMA) session where we delved into the exciting world of Atlassian Intelligence in Jira Service Management. With a particular focus on the innovative virtual agent, Atlassian product experts dove deep into the cutting-edge features and functionality we’ve been working on to enhance your service management capabilities. In this post, we highlight some of the key takeaways and top questions from the AMA.

Here’s our previous community post for the AMA and the event recording.


Key takeaways: 

Virtual Agent's Versatility in Customer Experience:

We showcased the virtual agent, offering seamless conversational support to your customers. Powered by Atlassian Intelligence, the virtual agent enhances customer experiences by providing intelligent answers and surfacing helpful documentation from your knowledge base.

Update on Microsoft Teams Integration

We’re committed to expanding our service management capabilities to more agents and help seekers, using Microsoft Teams. We’re targeting the second half of next year for the release of the virtual agent integration with Microsoft Teams. In the meantime, we recommend keeping an eye on our public cloud roadmap for updates.

Atlassian Intelligence Features in Early Access:

Our announcement of Atlassian Intelligence answers, issue summaries, and editor features (Early Access) generated a sizeable amount of excitement amongst attendees. These features significantly improve efficiency by allowing agents to summarize large volumes of comments on a ticket. You can also use Atlassian Intelligence in the issue editor to generate, transform, and summarize your writing. We can’t wait to see how these features help you work at high velocity!

On Data Privacy:

Atlassian does NOT send customer data to OpenAI’s platform to train its services. The data you submit and the responses you receive via Atlassian Intelligence are not used to fine-tune or improve OpenAI’s models or service. Each data request is sent to OpenAI individually, over an SSL-encrypted service, to process and send back to Atlassian.


Your questions 🤝 Our answers

Here are the questions from the event, including the ones we didn’t have time to answer! We’ll be adding more questions, answers, and resources to this post over the next few weeks.



Answers (to be updated)

Is the virtual agent going to be available to external / unlicensed users who want support from the JSM portal?

We plan to make the virtual agent available in other experiences in the future. While we can't provide a specific timeline, supporting unlicensed users is a possibility we're exploring.

Is there a JSM API to use Atlassian Intelligence programmatically?

Right now, there isn't a JSM API for using Atlassian Intelligence programmatically. However, we may consider adding including this functionality in a future roadmap.

Does AI answers come into the picture at all if the end-user's words trigger an intent with high confidence? Is it only for times where an intent can't be matched?

No, AI answers is only triggered if an intent is not matched with high confidence. The virtual agent initially attempts to identify an intent, and if a high-confidence match is not achieved, we then provide an AI response, assuming an answer is found. If neither is the case, the virtual agent asks the user to rephrase.

How can we control which Confluence spaces AI answers parses to get answers to end-user questions? 

You can control which Confluence spaces AI answers can parse by linking the desired spaces to your JSM projects.

Is there a plan to enable more Jira Automation that we can pick for out work in Jira and JSM?

We have lots of ideas on how to enable more automations! Currently, you can accomplish this with the Web Request Node. In the future, we hope to integrate with Jira Automations. Nothing we can definitively share yet.

Our instance of JSM does not show the "Summary" button. I do see the AI button in the editor, so it is activated. Is there any settings needed for making the "Summary" button show up?

AI summaries is available for all JSM Premium/Enterprise products. If you are not seeing the feature, please open a ticket with our support team.

Does the AI issue summary summarize all the comments on a ticket?

Yes, it summarizes all the comments, the summary, and the description on a ticket.

Will the Assist bot be customizable? It feels so robotic right now. The icon, name, tone etc. just screams "bot" when all our automated responses aim for a "human, fun, feeling"

You can customize the Bot name through your Slack settings. Some of the language can also be changed in the virtual agent settings. However, the icon is not currently changeable.

Will the virtual agent be available to Standard licensed users in the future?

At the moment, the virtual agent is only available on Premium and Enterprise plans. We currently have no plans to offer it on other tiers.

To clarify, only the AI portion of Assist is available on Slack, correct? I've seen Assist on Teams.

Correct, JSM Chat is available on MS Teams, but the virtual agent and AI answers are only available on Slack at this time. We recommend keeping an eye on our public cloud roadmap for updates and be sure to vote for the feature here.

So when it is searching Confluence, does that mean that OpenAI gets that content?

Please reference our AI trust page for more details on how we handle your data.

Can you have Confluence articles that are not visible to customers but are there to provide to AI answers?

At present, the articles must be non-private pages. We do not support having AI answers search articles not visible to customers.

When setting up multiple virtual agents, would those be tied to the same JSM?

You can connect different JSM sites to the same Slack workspace if you wish. Each virtual agent is connected to a single JSM project.

If we configure the virtual agent in Slack in the meantime, when it becomes available for Teams can we just transfer the agent over?

The configuration of the virtual agent in Slack will likely not be transferable to Teams when it becomes available.

Will intents ever be created automatically, based on previous replies?

Right now it feels like the opposite of AI / automation, as you need to manually set it up and create all the different choices it present.

Yes, if you have existing issues in JSM we recommend intent templates based on your past issues. You still need to configure the flows, but it already makes intent recommendations when you setup a new intent.

With AI answers pulling from Confluence, does it respect page restrictions? (Ex: in our Confluence space there are some pages that are restricted to specific teams that aren't for public consumption)

AI answers only pulls from non-private pages in Confluence. Currently, we do not support linking to other knowledge bases.

Can we link to other knowledge bases like Zendesk?

AI answers only pulls from non-private pages in Confluence. Currently, we do not support linking to other knowledge bases.

Is the mining of ticket data automatically setup? Or, is that something we need to request with Atlassian?

Intent templates are automatically set up based on the project’s past ticket data. The only requirement is to have enough ticket history to cluster the intents.

In "send message" is there the ability to add a picture? This would be immensely helpful in assisting our users.

In "send message", there is currently no ability to add a picture, but we've just created a feature request for you.

Does the virtual agent affect the automation monthly counts/limits in any way?

The virtual agent usage will be monitored separately from Automation usage.

In the flow of an intent, is there any intention to allow duplication of branches?

We do not currently allow duplication of branches in the flow of an intent. However, we appreciate your feedback and will consider it as we continue to improve the product. Here's a feature request you can vote on for this!

I remember recently to create a ticket via Slack you had to send an emoji to start the process. Is this something different or is this replacing that?

When the virtual agent is turned on, every message in the specified channel is responded to with the virtual agent.

If you are using JSM Chat without the virtual agent, you can configure the channel to still require a ticket emoji to turn into an issue.

Can the virtual agent be used to help categorize/manage requests submitted by email?

Currently, we do not support using the virtual agent to help categorize/manage requests submitted by email. Please follow this feature request.

Could we have a chat pop-up integrated to the JSM client webpage and not use Slack nor Teams?

Not at the moment. Please follow this feature request for updates.

Can the virtual agent be linked to assets?

Not at the moment. Here's a feature request for you to follow.

Will the virtual agent support multi-language?

The virtual agent allows you to customize your flows and copy in most places, but not completely. Feel free to vote and follow this feature request.

Do you plan on putting out any type of course within Atlassian University for setting up the virtual agent along with other useful tips on using it?

We are always looking for ways to support our users. While we don't currently have a course within Atlassian University specifically for setting up the virtual agent, we are considering various ways to provide more guidance and tips.

Will smart values be incorporated into the intents?

At the moment, smart values are not incorporated into the intents. However, we appreciate your feedback and will consider it for future updates.

Is the link within the Confluence page already clickable when presented by the virtual agent?

Yes, the links within the Confluence page are clickable when presented by the virtual agent.

Is the virtual agent accessible through Jira Service Management customer portals?

Currently, the virtual agent is not accessible through JSM customer portals. Please vote and follow this feature request to help us prioritize it

Can you use the AI summary feature to write an appropriate response?

The AI "summary" feature is designed to provide a concise overview of the issue and is not currently designed to write an appropriate response. We recommend using AI editor for that use case.

Is the summarize functionality based on the user?

The summarize functionality is not based on the user. It provides a summary of the issue content.

Do you have to continue to hit "summarize’ or will it keep up to date?

To keep the summary up to date, you need to click 'summarize' each time.

Can the summarize function be automated? (Included as an action in an automation)

The summarize function cannot be automated or included as an action in an automation.

Is the virtual agent part of the base product or is there additional cost?

The virtual agent is part of the base product for JSM Premium and Enterprise users.

Is there any interest in being able to call an automation from a flow? As well as using a scriptrunner script from a flow, also using automation to call the AI?

The web request step allows you to call any external API. This can be used to call an automation from a flow.

Does the summarization take from the last time you used the summarize action or from the creation of the issue?

The summarize function takes from the creation of the issue, not from the last time you used the summarize action.

Can you use the AI editor when creating Intents?

Currently, the AI editor cannot be used when creating Intents. We appreciate your feedback and will consider this for future updates.

Is that AI feature live? I can't find it to work the same way in my live environment. I see the snowflake but not with the tone, emojis, etc..

Many of our AI features are live in early access right now on Premium & Standard plans. If you're having trouble finding them, please refer to this link:

Can ticket content be turned into a knowledge base article?

At present, ticket content cannot be turned into a knowledge base article. Please follow this feature request for updates.

Is there self-learning by AI? Initially, we saw we could add natural language phrases but can it auto-populate those phrases based on last most reported usage?

The virtual agent recommends intents based on your past issues. It may use reinforcement learning to improve intents in the future, but does not currently.

I assume that tone improvement is only in Jira Service Management? Are there plans for Jira Software?

Tone improvement is available in Jira Software as well. You can find more information about the AI editor here:


Is it possible to get use of the sandbox before launching to our organization?

Yes, you can use the sandbox before launching to your organization. This is available on a per product/site basis at


Can the AI translate my comments?

The AI's ability to translate comments can vary. 

Is there any expectation to have Atlassian Intelligence generating insights or charts in JSM like it does for Atlassian Analytics?

Not currently, but we'll consider adding this feature in a future roadmap.

Can I use AI agent for only one portion of the functionality, for example, only use AI agent for summaries?

Currently, AI is all or nothing per product/site. But we'll consider adding more admin granularity in a future roadmap.

Is there any feature request to have native Atlassian chat...instead of relying on MS teams or Slack?

Not currently, but we’ll consider creating one just for this.

Can you tell the Atlassian Intelligence to reply back to the customer? Ex: “Write a response to let the customer know it will take a while to get this issue resolved”

Yes, you can instruct the AI to reply back to the customer, using AI editor. For example, you can say “Write a response to let the customer know it will take a while to get this issue resolved”.

Is there anything in the roadmap to generate a sentiment analysis of the ticket?

There is currently no feature on the roadmap to generate a sentiment analysis of the ticket.

Are there any new feature/ improvements coming out in the pipeline?

Yes, we have new features and improvements coming out in the pipeline. We recommend joining the AI waitlist.

Is the data extracted site wise, or at an Org level? We have 6 sites and would like to extract/train the AI based on 2 sites data.

The data extraction is project-based. You can choose to extract/train the AI based on specific site data.

Will assist be able to pull images / gifs into JSM tickets in the future? As some of our more techy employees have started interacting with Assist, they also send gifs... But sadly those are not added into the ticket (even though the ticket comments can send gifs normally)

Currently, Assist does not pull images / gifs into JSM tickets. However, we've created a feature request for you to follow!

How does the virtual agent know which request type to use in the project?

The virtual agent uses a default Request Type and offers an in-flow step to change request type within an intent flow.

Two quick questions: (a) Are any changes to the verbosity of the VA coming? (so many messages per thread, would be great to combine) (b) Is there any option coming for humans in support teams to “take over” a VA thread, stopping the flow when a human is best placed to respond?

Great feedback! The product is always evolving, we’ll share this with the product team.

Do you see any case where you would not recommend using the new AI functionality because of some potential risk involved?

We recommend that each organization evaluate the new AI functionality based on their specific needs and potential risks. For more information, please check out this page:

Is AI available in transition screens?

Not at the moment



Log in or Sign up to comment
AUG Leaders

Atlassian Community Events