Productboard is built around product management, but engineers are part of the product team. Spark, Productboard's AI agent, lets you read specs, ask product questions, and connect your own coding agent to Productboard directly. You don't need to wait on your product manager (PM) to translate.
This article covers how engineers and engineering managers get the most out of Productboard. It applies whether you're new to the platform or already using it alongside your PM.
Note: The Specification tab, Spark chat, and the Productboard MCP server described below require Spark to be enabled on your workspace. Every workspace has access to Spark, but activation is up to your admins. See Activating Spark in your workspace for details.
In this article:
- Reading and questioning specs without interrupting your PM
- Connecting your coding agent to Spark
- Keeping specs in sync with Jira and GitHub
- Staying aligned on priorities and dependencies
- What's next for engineers in Spark
- See also
Reading and questioning specs without interrupting your PM
Every initiative, feature, and subfeature has a Specification tab, which is a real-time collaborative editor where your PM drafts and refines the spec, often with Spark. Because it's real-time, you can open a spec while your PM is still editing it and see their changes as they happen.
If something in the spec is unclear, you don't have to interrupt your PM to find out. Instead, you can:
- Leave an inline comment directly on the section you have a question about, or type "@" followed by your PM's name to notify them.
- Open a Spark chat and ask your question directly. Spark reads the spec and related context and can answer many product questions on its own.
The Specification tab also keeps a full version history, so you can check what changed and when before assuming a requirement is final.
For the full walkthrough of writing, editing, and collaborating on specs, see Write product specifications with Spark.
Connecting your coding agent to Spark
Productboard's MCP server lets your coding agent read and act on specs directly from Productboard. This includes Claude Cowork, Claude Code, Cursor, and most other agents that support remote MCP connections.
A workspace admin must turn on MCP access for your workspace first. Once that's done, you connect your own agent to https://mcp.productboard.com and sign in with your existing Productboard account. There are no separate API keys to manage.
Once connected, your agent can:
- Find and read specs by name, owner, or tag, using your existing Productboard permissions.
- Propose edits to a spec, so decisions made mid-build get written back where the whole team can see them.
- Read and post comments, so a clarifying question reaches the spec owner instead of getting lost in Slack.
- Update status, reporting progress without a separate standup update.
Your agent works from the version of the spec your PM edited most recently, not an outdated copy. Treat any edit or comment your agent makes the same way you'd treat your own.
For full setup steps, see Connect a coding agent to Spark via MCP. You'll also find a technical guide covering the data-access model and how the server fits with spec-driven development frameworks.
Keeping specs in sync with Jira and GitHub
The Specification tab is a free-form drafting space and doesn't sync with delivery tools on its own. For that, Productboard's Details tab integrates with Jira and GitHub Issues. Push new issues or link existing ones to Productboard features for a two-way sync.
Once a spec in the Specification tab is ready, use the Fill from spec button in the item's Details tab. This copies it into the Description field, which then syncs with your connected delivery tool. Your day-to-day Jira or GitHub workflow stays unchanged, with a live, collaborative spec upstream of it.
Staying aligned on priorities and dependencies
Productboard's prioritization boards help your team focus on the features with the strongest evidence behind them. Features can be ranked by customer feedback and business goals, so engineering time goes toward what matters most.
Dependency tracking and shared roadmaps give engineering and product visibility into each other's plans without a dedicated status meeting. If your work depends on another team's feature, you can see that dependency on the roadmap. You don't have to wait for a meeting to find out about it.