Engineering Enablement by DX

Brian Scanlan is a Senior Principal Systems Engineer at Intercom, where he works on platform engineering, developer productivity, and AI adoption across the company.
In this session from DX Annual, Brian shares how Intercom set out to double engineering throughput and ultimately achieved that goal in nine months. Rather than treating AI as an optional productivity tool, the company standardized on Claude Code, updated performance expectations, invested heavily in enablement, and adopted an agent-first approach to technical work.
Brian explains why Intercom views Claude Code as a platform rather than a tool, how the company is building domain-specific skills and workflows for agents, and why it believes agents should eventually be able to perform any technical task a senior engineer can complete on a laptop.
He also shares the data behind Intercom's AI adoption efforts, including gains in throughput, reductions in defect backlogs, improvements in code quality, and the growing use of automated pull request approvals. Throughout the talk, Brian offers a practical look at what it takes to scale AI adoption across a large engineering organization and the lessons Intercom has learned along the way.

Where to find Brian Scanlan:
• LinkedIn: https://www.linkedin.com/in/scanlanb
• X: https://x.com/brian_scanlan 
• Website: https://brian.scanlan.ie

In this episode, we cover:
(00:00) Intro
(02:54) Intercom’s goal of doubling throughput 
(07:30) The platform strategy 
(09:30) Their agent-first strategy 
(10:58) Evergreen capabilities vs custom tooling 
(12:28) How Intercom works with agents
(16:43) What the data reveals about AI adoption and impact
(19:20) Using session data to improve AI workflows
(20:20) Cutting the defect backlog in half
(22:44) Inside Intercom’s Claude Code setup
(28:09) Claude Code beyond engineering
(30:49) Q&A #1: Token cost 
(32:52) Q&A #2: Preparing for AI pricing changes
(34:14) Q&A #3: Stress testing and auditing skills
(36:31) Q&A #4: Criteria for agents approving PRs

Referenced:
• Intercom
• Software? No Way. We’re an A.I. Company Now! - The New York Times
• Anthropic
• Snowflake
• Linear
• LaunchDarkly 
• Fin AI
• Microsoft Copilot
• Cursor
• Claude Code | Anthropic's agentic coding system
• Steve Yegge (@Steve_Yegge) / Posts / X 
• Honeycomb
• Fin Ideas
• Fin CLI | AI Agent Command Line Interface

Michael Galloway leads Platform Engineering at Mercari, while Snehal Shinde leads Cost and Performance Engineering. Together, they have been at the center of Mercari's effort to become an AI-native company.
In this session from DX Annual, Michael and Snehal share what happened after Mercari's CEO mandated 100% AI adoption across the organization. While AI accelerated code generation and increased engineering output, the team quickly discovered that their existing dashboards could not answer a simple question: was AI actually improving productivity?
They discuss how Mercari built new visibility into AI usage and software delivery, the bottlenecks they uncovered across the SDLC, why faster coding did not automatically translate into faster delivery, and the lessons they learned rolling out AI at scale. They also share how Mercari is rethinking software development around agents, feedback loops, and new ways of working.

In this episode, we cover:
(00:00) Intro
(01:46) Mercari’s scale and engineering culture
(02:51) DX awards at Mercari
(03:44) Mercari’s push to become AI-native
(06:34) The mandate to rethink everything
(08:02) Mercari’s AI visibility problem and how they solved it
(11:30) Mercari’s early findings on AI implementation
(18:47) Closing the AI awareness gap at Mercari
(21:11) Mapping AI opportunities across Mercari
(31:32) Unpacking the results from the second rollout
(34:14) Agent spec-driven development and what’s next
(37:37) A multi-loop SDLC
(40:50) Some hard lessons
(42:55) Closing thoughts

Referenced:
• Mercari
• Cursor
• Devin
• Claude Code | Anthropic's agentic coding system
• GitHub
• Datadog
• Tim Bozarth - Microsoft | LinkedIn
• Airbnb
• Jim Collins - Concepts - The Stockdale Paradox

In this session from DX Annual, Eleanor Millman, Senior Staff Product Manager, and Mina Tawadrous, Associate Director of Product Management at SiriusXM, share how their platform engineering organization developed a prioritization framework for platform engineering teams serving hundreds of developers across a complex cloud platform.
They explain how they define and weight platform-specific impact factors, use developer data to refine priorities, and score projects more consistently. They also explore why prioritization debates often stem from conflicting, invisible, or outdated assumptions, and how SiriusXM began treating assumptions like code by documenting, versioning, and reviewing them in source control.
Finally, they demonstrate how AI can surface assumptions, connect initiatives to existing knowledge, and support project scoring while keeping humans in the loop. Throughout the session, they offer a practical framework for making prioritization decisions more transparent, data-driven, and scalable.

In this episode, we cover:
(00:00) Intro
(02:58) Building a platform engineering prioritization framework
(04:59) The seven platform engineering impact factors
(09:38) Using impact factors to score projects
(13:11) Using developer data to refine priorities
(16:33) Three ways assumptions fail 
(17:40) Assumptions as code 
(21:00) New problems created by assumptions as code
(22:00) Using AI to surface assumptions
(23:44) Building an AI-powered feedback loop
(25:44) Inside the AI prioritization tool
(28:18) Three steps to build your own framework
(30:02) Q&A #1: Evaluating high-cost projects
(31:30) Q&A #2: The cadence of iteration 
(32:10) Q&A #3: When the framework conflicts with a stakeholder's priorities
(35:26) Q&A #4: Using the framework for non-developers

Referenced:
• AWS
• Databricks
• RICE: Simple prioritization for product managers
• Designing developer experience surveys
• GSB Preserve | View | The Curse of Knowledge

Kelly Anne Pipe is Head of Developer Experience at Vanguard, and Nicole Scribner is a Director in the firm's Chief Technology Office focused on engineering enablement and advancement.
In this session from DX Annual, Kelly Anne and Nicole share how Vanguard is expanding its AI strategy beyond software engineering to the entire product development lifecycle. While the company initially focused on tools like GitHub Copilot for engineers, they found that faster coding alone did not significantly improve delivery speed. Product managers, designers, QA teams, and organizational processes were still operating at a different pace.
To address this challenge, Vanguard developed a product team maturity model built around three stages: Augmented, Accelerated, and Autonomized. The framework spans six dimensions, from AI-powered delivery and AI-ready codebases to team autonomy, operations, and responsible AI.
Kelly Anne and Nicole explain how Vanguard is applying the model across more than 800 product teams, the behaviors they believe will enable faster delivery, and the lessons they have learned about measurement, organizational change, dependencies, and scaling AI across the product development lifecycle.

In this episode, we cover:
(00:00) Intro
(02:16) The state of AI one year ago at Vanguard
(02:54) The engineering bubble
(05:05) Building an AI maturity model for 800 product teams
(08:24) Dimension 1: AI-powered product delivery
(10:00) Dimension 2: AI-ready codebase
(12:20) Dimension 3: Autonomous agent utilization 
(13:00) Dimension 4: AI-augmented operations
(14:00) Dimension 5: Team autonomy and enablement
(16:11) Dimension 6: Responsible AI
(18:15) The people problem: role evolution 
(20:00) The measurement problem 
(22:55) Lessons learned from rolling out the maturity model 
(26:46) What’s ahead 
(30:10) Q&A #1: Getting your codebase ready for AI
(32:22) Q&A #2: Audit trails and responsible AI
(34:16) Q&A #3: Vanguard's maturity model progress
(36:15) Q&A #4: Measuring cycle time across 800 teams

Referenced:
• Vanguard
• Jennifer St Pierre - Dell Technologies | LinkedIn
• Mercari

Recorded live at DX Annual, Abi Noda, co-founder and CEO of DX, joins Brian Houck of Microsoft to share an early look at DX’s new research on AI’s impact on engineering velocity.
Drawing on data from a sample of DX customers, they discuss what companies are actually seeing as AI adoption matures. Most organizations in the study saw pull request throughput increase by 10 to 15 percent—far more modest than the 10x gains often promised in industry headlines.
They explore why coding remains only a small part of developer work, where time saved by AI may be going, and the unintended consequences of moving faster, from shifting bottlenecks to “false velocity.” Abi also shares how engineering leaders are applying AI beyond coding and how DX is evolving its measurement framework to account for both human and agent productivity.

Where to find Brian Houck: 
• LinkedIn: https://www.linkedin.com/in/brianhouck/ 

Where to find Abi Noda:
• LinkedIn: https://www.linkedin.com/in/abinoda

In this episode, we cover:
(00:00) Intro
(00:53) What motivated DX’s research into AI’s impact on engineering velocity
(02:36) How DX designed the study and selected companies
(04:54) What DX’s data reveals about AI’s impact on engineering throughput
(06:31) Why PR throughput was the most practical metric to publish
(08:21) Why AI productivity gains are lower than many leaders expected
(10:24) How an all-in culture can amplify AI productivity gains
(12:35) Why it’s hard to track where AI-generated time savings are going
(15:04) Unintended consequences of AI-driven productivity gains
(17:12) Why leaders should look beyond coding to the rest of the SDLC
(19:43) Cognitive debt and the human costs of AI-assisted development
(21:33) How DX’s AI measurement framework is evolving
(24:42) How to make agents more effective

Referenced:
• DX Core 4 Productivity Framework 
• DORA, SPACE, and DevEx: Which framework should you use?
• Time Warp: The Gap Between Developers’ Ideal vs Actual Workweeks in an AI-Driven Era - Microsoft • Research
• How Generative and Agentic AI Shift Concern from Technical Debt to Cognitive Debt
• Measuring AI code assistants and agents