Platform bundling
For an organization fully committed to Bitbucket, Jira, and Confluence, a native, bundled AI measurement layer is a real convenience.
Tetriz vs DX
DX blends measured data with a self-reported input. Tetriz measures the session directly.
Atlassian's Team '26 announcements (AI Code Insights, AI Pulse, and the Rovo Dev coding agent) are genuine, shipped capabilities bundled into Jira and Bitbucket. Tetriz reads the AI coding session itself by default, independent of any single vendor's stack or its own coding agent.
A measurement platform that also sells the agent it measures
DX's research frameworks, DXI and DX Core 4, were developed with the researchers behind SPACE and DORA, and its customer and conference base is large and credible. Those are genuine strengths, and Rovo Dev, Atlassian's own coding agent, has ranked at the top of the SWE-bench leaderboard. The issue is structural: the same company measuring how well an engineering org directs AI now also sells one of the agents being measured. What actually decides whether an org is becoming AI-native sits underneath that figure: the session itself, captured by default, scored across six prompt-quality dimensions the engineer can act on directly. Tetriz has no coding agent to sell, and reads the session directly across every agent it supports.
Where Tetriz wins vs DX. Session-level data, mapped to the pull request it produced.
| Capability | Tetriz | DX |
|---|---|---|
| Session-level capture | Reads the actual IDE session via the Tetriz Desktop App by default, and scores six prompt-quality dimensions. | AI Code Insights can capture session content, including prompts, for natively-hooked tools, but only as an optional feature an admin must configure. Worth verifying what's enabled in a given deployment. |
| Identity attribution | A single verified identity anchor from account sign-in. | Resolves identity through a git-config and commit-email fallback chain. Worth asking DX directly how it behaves when those signals disagree. |
| AI-impact methodology | Session time and pull-request attribution, measured directly from the Tetriz Desktop App, with no self-report step. | The published 'Net AI Dollar Impact' formula sums four components. One of them, the efficiency-gain input, comes from a self-reported survey. |
| Vendor neutrality | Six supported coding agents, with no coding agent of its own in the products it scores. | Now also sells Rovo Dev, a coding agent, alongside the platform that scores how engineers direct AI tools generally. |
Where DX wins. Said plainly, credit where it’s due.
Research pedigree & Rovo Dev
DXI and DX Core 4 remain among the most credible developer-experience research frameworks in the industry, and Rovo Dev is a genuinely capable coding agent in its own right. Tetriz measures how well any agent, including Rovo Dev, is being directed; it doesn't build developer-experience research or a coding agent.
Questions teams ask. Comparing Tetriz and DX.
For tools it hooks into natively (Cursor, Claude Code, Copilot), it can capture full session data, including prompt content, but only as an optional, admin-configured feature, not the default. For other tools, it relies on commit and file-change heuristics instead. Worth confirming with DX which mode is live for a given org before assuming parity with a default-on, IDE-native capture.
Atlassian now sells Rovo Dev, a coding agent, alongside the platform measuring how well engineers direct AI tools generally, including Rovo Dev itself. That's a structural overlap worth being aware of when reading its numbers. Tetriz doesn't build a coding agent, so that particular overlap doesn't apply to it.