# Rovo at Work: AI agents in action Source: https://www.youtube.com/watch?v=0PId38C4kfg Recap page: https://rapidrecap.app/video/0PId38C4kfg Generated: 2026-02-25T19:27:18.753+00:00 --- ## Quick Overview Rovo agents, deeply embedded within the Atlassian stack (Jira, Confluence, etc.) and enhanced by the Rovo MCP Gallery, allow teams to automate complex, multi-step workflows directly through natural language chat, saving time on manual tasks and enabling AI to act as a true teammate across planning, development, and documentation. **Key Points:** - Rovo agents are deeply embedded within the Atlassian stack (Jira, Confluence) and can execute tasks directly within those applications. - The Rovo MCP Gallery allows teams to connect agents with powerful third-party skills (like Amplitude, Canva, Figma) to extend their capabilities beyond Atlassian products. - The process demonstrated involved an 'Email Writer' agent drafting a recruitment email based on existing Jira issue context and then a 'GitHub Copilot' agent reviewing the associated pull request, all automated via workflow triggers. - Rovo Studio allows users to build custom agents using natural language prompts, defining their role, instructions, and the specific skills (including MCP skills) they should use. - The goal of these AI agents is to eliminate tedious manual work, such as context-switching between tools, allowing teams to focus on higher-value, strategic work. - The platform emphasizes agent accountability by clearly showing which agent is working on a task and providing visibility into the process, ensuring humans remain in control. - New features include 'Rovo Studio' (for building agents) and 'Rovo MCP Gallery' (for third-party skills), both available in early access. ![Screenshot at 00:08: Gunjan Sood introduces the focus on practical AI agents embedded within Jira, Confluence, and the rest of the Atlassian stack, designed to act like a teammate.](https://ss.rapidrecap.app/screens/0PId38C4kfg/00-00-08.jpg) **Context:** This video, titled 'Rovo at Work: AI agents in action,' features Gunjan Sood (Head of Product Management, AI, Atlassian), Alfredo Huitron (Product Manager, AI Agents, Atlassian), Melissa Miller (Head of Product Marketing, AI, Atlassian), and Matthew Canham (Principal Product Manager, AI, Atlassian) discussing the evolution and practical application of Atlassian's AI agents, particularly focusing on the new Rovo Studio and Rovo MCP Gallery capabilities. ## Detailed Analysis The presentation highlights how Rovo AI agents move beyond theoretical discussions to provide practical, integrated automation within the Atlassian ecosystem. Gunjan Sood emphasizes that agents can work directly inside Jira, Confluence, and the rest of the stack, saving teams time on manual tasks and allowing focus on strategic work. Alfredo Huitron details the pain point of context switching between tools like Figma, Amplitude, and Canva, which slows down work. He explains that Rovo Studio allows users to build custom agents from scratch or via templates, using natural language to define logic, such as creating a bi-weekly feature prioritization brief that synthesizes Jira data with market insights. Matthew Canham further explains that agents are assignable within Jira tickets, providing accountability, and can be triggered automatically when work moves through workflow stages (e.g., moving a ticket to 'In Progress' can trigger a GitHub Copilot agent to check linked repositories and PRs). Melissa Miller introduces the Rovo MCP Gallery, which allows agents to connect to thousands of third-party services (Amplitude, Canva, etc.) via skills, enabling them to perform complex actions like generating branded graphics in Canva or creating presentations in Google Slides, all without the user leaving Jira or breaking flow. The core theme is that AI becomes the central orchestration layer, seamlessly connecting tools and processes to increase team velocity and decision-making quality. ### Introduction of Rovo Agents - Gunjan Sood introduces the focus on practical AI integration across the Atlassian stack - Alfredo Huitron highlights the problem of context switching between tools like Figma and Canva - Matthew Canham sets the stage for demonstrating agent collaboration within Jira. ### Agent Building with Rovo Studio - Alfredo demonstrates building a 'Feature Prioritization Analyst' agent using natural language prompts and specifying assumptions for bi-weekly execution - The agent's steps involve analyzing Jira data and creating a Confluence page. ### Agent Capabilities via MCP Gallery - Melissa Miller explains that Rovo agents extend beyond Atlassian tools by connecting to third-party MCP servers (Amplitude, Canva, etc.) via skills - The 'Blog Creator Agent' example shows adding a Canva skill to generate on-brand visual assets for blog posts. ### Agent Action in Jira Workflows - Matthew demonstrates assigning the 'Email Writer' agent to a Jira ticket to draft a recruitment email, which then posts the draft directly to the work item's comments - Subsequently, the 'GitHub Copilot' agent is assigned to review the linked PR, inspect code, and provide suggestions directly in the Jira comments. ### Conclusion on AI Collaboration - The team concludes that AI agents become the central orchestration layer, enabling seamless handoffs between humans and AI across different tools (Jira, Confluence, GitHub) while maintaining accountability and control. ![Screenshot at 00:08: Gunjan Sood introducing the concept of practical AI agents working inside Jira and Confluence.](https://ss.rapidrecap.app/screens/0PId38C4kfg/00-00-08.jpg) ![Screenshot at 02:09: The Rovo Studio interface where a user defines an agent's purpose using natural language prompts.](https://ss.rapidrecap.app/screens/0PId38C4kfg/00-02-09.jpg) ![Screenshot at 06:42: A demonstration showing a custom agent's instructions being refined in Rovo Studio, including the use of a Canva skill.](https://ss.rapidrecap.app/screens/0PId38C4kfg/00-06-42.jpg) ![Screenshot at 11:01: Demonstration of an 'Email Writer' agent drafting content directly into a Jira ticket's comment section.](https://ss.rapidrecap.app/screens/0PId38C4kfg/00-11-01.jpg) ![Screenshot at 12:33: A Jira board view showing how an agent's work \(GitHub Copilot\) is automatically triggered when a ticket moves to the 'In Progress' column.](https://ss.rapidrecap.app/screens/0PId38C4kfg/00-12-33.jpg)