Atlassian · AI product · Mid · 45 to 60 minutes

Where should Rovo MCP invest next?

Atlassian makes collaboration and work management software used by software and business teams. Answer this product management case and get it graded line by line against a weighted rubric.

Segmentation
Product strategy
Prioritization
Metrics
AI product judgment

The situation

You are the PM for Atlassian's Rovo MCP server, which lets AI agents access work context and act across Atlassian products and connected tools. Usage is growing quickly enough that leadership wants a sharper view of who this product is really for and what should be built next. Some teams argue MCP is mainly a technical integration surface and should focus on reliability, deeper developer workflows, and more actions for agents. Others think the larger opportunity may be people whose jobs depend on gathering context, synthesizing updates, and coordinating across systems, but the current product and onboarding are not designed with those users in mind. The VP of Product asks you for a recommendation before next quarter planning. You need to decide how to segment users, what signals would prove real product value, and which roadmap bets deserve the next two quarters of investment. The hard part is that aggregate agent activity can look strong while hiding who gets repeat value, and a wrong focus could either slow adoption or build depth for the wrong audience.

What to produce

A 1 to 2 page product strategy memo.

Your answer must

State your primary recommendation for the next two quarters, and explain the main trade-offs behind it.
Define 2 to 3 target user segments with job to be done hypotheses, including how you would test whether non-technical roles are worth serving now.
Propose a metrics framework with leading and lagging indicators, including how you would identify power users and repeated useful agent work.
Recommend the top 3 roadmap bets and 1 clear deprioritized area, with rationale tied to user value and enterprise fit.
Outline a 90 day learning plan that specifies what instrumentation or experiments you would run and what findings would change your mind.

Constraints

Must build on Atlassian's existing products and the Rovo MCP server, not a brand new standalone product.
Recommendations must work for enterprise customers with strong expectations around privacy, permissions, and admin control.
Assume the team can pursue only three meaningful roadmap bets over the next two quarters.

Format

Use headings: decision, target segments, metrics, roadmap, risks and open questions.

How Atlassian actually played it

Atlassian shares what usage of its Rovo MCP server shows about how agents are really used at scale, including more than five million tool calls on each working day. One surprise is that 44 percent of MCP users are not on software teams, and executives and PMs use it as heavily as engineers. A good example of learning who your power users really are from usage data.

Check who is actually using a developer-oriented interface, because the segment may be wider than planned.
People whose job is synthesis across tools value an agent connection most.
A shared context layer is the reason agents can do useful work in a product.
Publish usage findings to shape your own roadmap and the market's view.

Built from What 5M+ daily MCP tool calls taught us about the future of AI at work by Sherif Mansour, Atlassian.