|
How do you bring leadership on the journey of the benefits of prioritising the use of AI on an integrated way to all of the key tools in the business. Do you have study cases you can share to be used in our organisations?
The best way to engage leadership in AI is to link it to their priorities—speed, cost, quality, risk—and demonstrate value through small, measurable steps.
With Rovo, customers typically:
-
Start with 1–2 clear use cases (e.g., using Rovo to search across Jira, Confluence, and other tools, or automating routine updates with agents). If you face a daily pain point, Rovo can help. Begin with the problem you want to solve.
-
Highlight the benefit of an integrated AI layer across tools instead of many disconnected bots, simplifying governance, scaling, and reporting on value.
-
Prioritize trust and governance early by showing how Rovo respects permissions, reveals answer sources, and keeps company data secure.
In this case study Mary's Meals shares a few of their Rovo use cases.
|
|
Where would be the best place to start for a company to use Rovo? Would it be admin tasks, or project updates as demoed today - or would you start somewhere else?
The easiest place to start is with Rovo Search. This acts as an AI supercharged way to find information. It is an action that people are already familiar with and people immediately see the value - saving time by getting to the right information faster. What will be crucial is connecting your data that your team uses most to make sure people can find and chat with not only Confluence and Jira but any other of the tools you use in one place.
The next step can be to find a couple ways to embed Rovo into regular workflows - for EOF that was their weekly project updates as they demoed. For Taylor that was directing his team when they ask a question to ask Rovo Chat first.
Paul also thinks that the best place to start w Rovo is organisation unstructured data, like updates or even meeting notes via loom. With minimum lift these can become a huge resource and instant IP, where previously they were hard to utilise.
|
|
How do you ensure the output of Rovo is accurate? Is it missing something critical? Is it hallucinating? Is it making up information?
Rovo is built to provide responses grounded in your organization’s actual data like Jira issues, Confluence pages, and other connected sources using Atlassian’s Teamwork Graph. This means answers are based on the information you and your team have access to, not just generic internet knowledge. Rovo also has checks to try to minimize hallucination.
Rovo provides source transparency so you can always see which sources were used for an answer.
Rovo is built to maximize accuracy and transparency, but with all AI, it’s important to lean into human-AI collaboration to check it. Users should always review answers, check citations, and provide feedback to help Rovo continuously improve.
|
|
How does Rovo technically know organizational data over Atlassian tools? Reason behind this: Sometimes we get general answers from Rovo, even we set the scope for only one Jira project or Confluence space.
Rovo combines two things when it answers:
-
Your organization’s data
-
Jira (issues, comments, fields, etc.)
-
Confluence (pages, comments, attachments metadata)
-
Jira Service Management, Jira Product Discovery
-
Connected third‑party apps (e.g. Google Drive, Slack, custom websites) via Rovo Search
-
Anything agents are explicitly allowed to read or act on
-
General product / domain knowledge Beyond “how to use Atlassian tools,” Rovo’s underlying models have broad world and work knowledge, for example:
-
General software & agile practices
-
Concepts like user stories, epics, sprints, incident management, SLAs, OKRs, etc.
-
Best‑practice patterns (e.g. how to run retros, how to structure a PRD, how to triage incidents).
-
Generic writing / communication skills
-
Drafting summaries, emails, announcements, specs, FAQs, meeting notes.
-
Rewriting, shortening, changing tone, translating, etc.
-
Generic reasoning and analysis
-
Basic cross‑tool workflows (non‑Atlassian specific)
So when you see an answer that feels “general” even though you’ve scoped to a project/space, it’s usually because:
-
Your question is broad / best‑practice‑y (“How should we manage sprints?”) so the model leans on its general agile knowledge.
-
Or there’s not enough rich content in that specific project/space, so there’s limited organization‑specific material to ground the answer on, even though the search was scoped.
You can often nudge Rovo away from generic knowledge by:
|
|
I’m eager to use AI, but I’m concerned about accuracy. Does Rovo always cite its sources—like specific Jira tickets or Confluence pages? I want to be able to verify the information before I act on it.
Yes, Rovo prioritizes citing its sources at the bottom of the answer and also provides links so users can verify the details for every response.
The sources cited will be from any tools connected to Rovo including all Atlassian apps as well as connected third party apps. For example in one answer I may see sources cited as a Confluence page, Google Sheet, Slack message, and Jira ticket.
If it doesn’t, we encourage asking Rovo Chat for sources directly.
|