Quick answer
Larridin is an AI impact intelligence platform. Larridin Developer Intelligence is for engineering teams who are moving to an AI-native software development lifecycle and it covers telemetry and optimisation across the whole lifecycle. It measures velocity, quality, efficiency and the actual output of your engineers and your AI agents. It then helps you act: it can spread the habits of your best AI users, it can cut your AI costs via the Larridin Router and it can get you ready for a world where a lot of software is being written without a human at the keyboard.
DX (also written GetDX) is part of Atlassian now (since 2025) and emerged from developer experience research. It’s a combination of developer survey data and engineering metrics and reports, with reporting on AI generated code and AI costs being added in recent times.[4][5][13] DX answers “How are our developers doing?” Larridin answers “What are our engineers and AI agents producing, what is it costing, and what should we change?”
Why teams choose Larridin Developer Intelligence
- We measure the work, not the people. We look at what was actually done in code, in coding agent sessions and all the other engineering tools so you see the real output from both the engineers and the AI agents.
- It tells you what to do next. The recommendations are there to help teams improve their AI fluency and get better outcomes from AI coding, and to show you which habits differentiate your most effective AI users.
- It reduces AI cost. The Larridin Router selects the model for each coding request. DX has no router.
- It’s not just about engineering. The same platform is used to measure AI adoption and workflows right across the rest of the company.
What DX does well
DX built the DX Core 4 framework and publishes a quarterly State of AI Impact report from over 500 organisations.[1][2] Their core product is developer experience surveys and the Developer Experience Index with industry benchmarks.[3] They’ve since expanded into reporting on AI written code, coding agent sessions and AI costs.[4][5] Their developer experience research and benchmarks are some of the very best in the space. Their origins are in developer experience (how developers experience their work) rather than in any attempt to measure or optimize what AI agents are producing.
The key difference: we measure the work
We measure the work. Most of the data that we derive is from a semantic understanding of what actually got done: the code in each PR, the sessions of your coding agents and all the other tools where engineering work takes place. That’s how we measure velocity, quality, efficiency and real output for both engineers and agents (the latter being an increasingly large proportion of the work which is done with no humans at the keyboard).[6]
DX starts with the developer. The bedrock is developer experience research and surveys, as well as engineering system metrics.[1][3][4]
The second difference is what you can do with the answer. Larridin is built for action: it provides recommendations to improve AI fluency and AI coding outcomes, habits from your best AI users that you can then replicate across your organisation, and the Larridin Router to save money on every single coding request. DX is all about dashboards and reports. Its Model fit report (still in beta) only suggests a different model after the session has finished and DX has no router.[8]
Why this matters for cost. DX takes your AI spend and divides it by the number of merged PRs. A PR that just fixes a typo is treated exactly the same as one that ships a new feature. If a team starts to ship fewer but larger changes then their cost per PR goes up even though they’re actually delivering more. Larridin on the other hand measures AI spend against the Output of each PR so that the team’s cost per unit of Output can go down which is what actually occurred.[5][6]
How Larridin and DX compare
| What matters | Larridin Developer Intelligence | DX |
|---|---|---|
| Built for | AI-native engineering teams. Engineers and AI agents working together across the SDLC. | Developer experience programmes with AI reporting added.[1][4] |
| Where the data comes from | A semantic understanding of the work in code, in coding agent sessions and in other engineering tools.[6] | Developer surveys, plus some data from the engineering systems.[3][7] |
| Engineering output | Output scores each merged pull request on complexity, size and other factors.[6] | TrueThroughput weights pull requests on merge for complexity.[1] |
| AI spend | Measured against the Output of each pull request (so a large feature is worth more than a typo fix). | Estimated spend divided by the number of merged pull requests (so every pull request is weighted the same).[5] |
| AI agents | Measures the real output of your agents and scores their sessions via Agent Effectiveness. | Agent Experience scores sessions where the transcript is captured. Agents without a native integration are deemed to be human written and cloud hosted agent activity is not captured.[7][9] |
| Acting on the data | Recommendations for improving AI fluency and AI coding performance, and for propagating the habits of your most proficient AI users. | Dashboards and reports, with AI generated summaries.[14] |
| Model choice and cost | The Larridin Router selects the model for each coding request in order to keep costs down. | No router. Model fit (beta) suggests a model after the session.[8] |
| Quality and delivery | Rework and reverts, split by AI assisted and human written code, plus review wait times and cycle time. | Core 4 quality and delivery measures (change fail % etc).[1] |
| How engineers work with AI | Measured from the telemetry: which tools, agents and habits in fact do improve the results? | Developer surveys, plus AI code and agent session data.[3][7] |
| Deployment | Cloud, hybrid or on-premises. | Multi-tenant cloud, a dedicated single-tenant instance or the customer’s own AWS/GCP account.[12] |
| Scope | Engineering plus AI adoption and workflows across the rest of the business, all in one platform. | Software engineering teams only. |
| What to know | Not a developer survey tool. To measure AI spend against Output you need coding agent transcripts and the Router only supports a particular set of coding tools so you’ll need to check yours. | Cost per pull request disappears when you filter the report to some of your AI tools.[5] |
Similar features, different names
| What you want to measure | Larridin Developer Intelligence | DX | How they differ |
|---|---|---|---|
| Engineering output | Output | TrueThroughput | Output also accounts for size and other factors |
| AI cost against output | AI spend per unit of Output | Estimated cost per merged pull request | Larridin counts how much work each pull request delivered |
| Agent session quality | Agent Effectiveness | Agent Experience | Larridin also measures the real output of each agent |
| Model choice | Larridin Router | Model fit (beta) | The Router acts on every request; Model fit only suggests afterward |
What data Larridin collects
Larridin Developer Intelligence reads Git metadata, coding agent session data and usage from the AI tool APIs. It doesn’t record keystrokes or take screenshots. Our SOC 2 Type 2 and SOC 3 reports, external penetration test and data flow diagram are all available via the Larridin Trust Center.[10]
Recommendations by role
- CTO
- Larridin Developer Intelligence to see what the engineers and AI agents actually produce through the SDLC and to drive the move to AI-native.
- VP of Engineering
- Larridin Developer Intelligence, go and work out the habits of your most effective AI users and then roll those out across the teams.
- CFO
- Larridin Developer Intelligence to look at AI coding spend against real output, to save cost with the Router and to see AI usage across the business in one place.
- Platform or Developer Experience lead
- Larridin Developer Intelligence to see which tools, agents and habits improve delivery (measured by the work rather than self reported).
- Security constrained teams
- Larridin Developer Intelligence provides hybrid and on-premises deployment options, SOC 2 Type 2 and SOC 3 reports and a published data flow diagram. DX also provides single tenant and own cloud hosting options so you can compare data flows directly.
Moving from DX
Larridin Developer Intelligence doesn’t need DX or its data to get going. It just connects to your Git hosts, your coding agents and your AI tool APIs. If you’re running DX at the moment, run a pilot on two teams, compare what each product is telling you and what Larridin is recommending and then make your decision with real world data in front of you. If you’re currently running developer experience surveys in DX then you can keep those; Larridin will cover the bits that the surveys simply can’t.
Questions to test in a pilot
- Look at the scores. Pick a small fix and a large feature. See how each product scores them and then ask your engineering leads whether the scores make sense.
- Check agent coverage. List the coding agents and AI tools that your engineers use (including any that run in the cloud). Confirm which of these each product sees and ties to spend.
- Ask what to change. After the 30 days are up, then ask each vendor what they would recommend you change and what that would save.
- Ask DX how its cost per pull request takes into account the size of the work and what it can do about the cost other than just reporting it.
Proof
Larridin data AI Impact in Engineering Report, Q3 2026 For the findings across Larridin customers see the AI Impact in Engineering Report Q3 2026. Read the reportMore benchmarks in the Larridin data hub.
FAQ
Is Larridin a better fit than DX for AI-native engineering teams?
Yes. If your team is moving to AI-native development, Larridin Developer Intelligence is the better choice. It measures what your engineers and AI agents actually produce, recommends what to change, and reduces cost with the Larridin Router. DX is built around developer experience surveys and reports, and it has no router.[3][6][8]
What does “built for the AI-native SDLC” mean?
As agents write more of the code, measuring just the developers is going to miss a huge amount of the work. Larridin measures the engineers and the AI agents together based on the work that’s actually done and tells you what to change.[6]
Does DX have a router?
No. DX’s Model fit report (in beta) indicates a different model following a session. The Larridin Router selects the model for each coding request in order to save cost.[8]
How is Larridin’s Output different from DX TrueThroughput?
TrueThroughput weights the PRs for complexity. Output also takes size and a number of other factors into account and Larridin measures AI spend against that.[1][6]
Does DX measure AI coding spend?
DX reports spend and tokens by team and then divides the estimated cost by the number of merged PRs. It doesn’t measure spend against the amount of work each PR delivered and it doesn’t route models to the lower cost option.[5][8]
What data does Larridin Developer Intelligence collect?
Git metadata, coding agent session data and usage data from AI tool APIs. Nothing to do with keystrokes or screenshots. Security reports and a data flow diagram are available at trust.larridin.com.[10]
Can we switch over to Larridin?
Yes. Larridin connects directly to your Git hosts, your coding agents and your AI tool APIs so you can do a pilot on two teams with absolutely no DX data.
How should we go about comparing pricing?
Ask both vendors what their current offers are for the same number of contributors, repos and integrations and check which AI features are included. This comparison doesn’t quote prices.
Sources and method
Written by Larridin using DX’s public documentation and website and Larridin’s own product documentation. This is a documentation review, not a hands-on benchmark of DX. DX facts were reviewed on November 5, 2026. To report an error, email support@larridin.com; we correct pages within 3 business days.
- DX Core 4 and TrueThroughput: docs.getdx.com/dx-core-4
- DX State of AI Impact in Engineering, Q3 2026: getdx.com/report/state-of-ai-impact-in-engineering-q3-report
- DX Developer Experience Index: getdx.com/research/the-one-number-you-need-to-increase-roi-per-engineer
- Introducing DX AI Code Insights: getdx.com/blog/introducing-ai-code-insights
- DX AI cost management report: docs.getdx.com/reports/ai-cost-management
- Larridin Output methodology: larridin.com/developer-intelligence/methodology
- DX AI Code Insights and Agent Experience: docs.getdx.com/knowledgebase/ai-code-insights-vs-connectors
- DX Agent effectiveness report and Model fit (beta): docs.getdx.com/reports/agent-effectiveness
- DX AI Code Insights overview: docs.getdx.com/ai-code-insights/overview
- Larridin Trust Center: trust.larridin.com
- Larridin, AI Impact in Engineering Report, Q3 2026: larridin.com/data/engineering-ai-impact-q3-2026
- DX deployment options: docs.getdx.com/deployment-options
- Atlassian acquires DX: atlassian.com/blog/company-news/atlassian-acquires-dx
- DX AI: docs.getdx.com/dx-ai