From Scattered Signals to Organizational Intelligence: AI-Powered Knowledge Capture at Scale

Your team is learning constantly — but most of that knowledge evaporates. Learn practical patterns for capturing and surfacing organizational intelligence, using AI to amplify both what goes in and what comes out, without adding burden to busy engineers.

Every week, your engineers, tech leads, and operations teams are encountering patterns — repeated friction points, emerging risks, quiet wins, and hard-won lessons. Most of it never gets written down. Some of it ends up in a retro that nobody reads. Almost none of it reaches the people who need it most.

This talk is about closing that gap, using AI not as a replacement for human judgment, but as infrastructure for knowledge capture and synthesis at a scale that was previously impossible.

For the past 18 months, I’ve led the team behind the Thoughtworks Technology Radar — one of the most recognized knowledge-sharing artifacts in the industry. That experience gave me a front-row seat to what makes organizational learning actually work: the discipline of capturing signals across a large, distributed organization, synthesizing them into something coherent and opinionated, and presenting them in a way that different audiences can act on. It also showed me where even well-resourced, highly intentional processes break down — and where AI changes the equation.

This talk translates those lessons into patterns any organization can adopt. We’ll look at how to design lightweight input mechanisms that respect engineers’ time (think: 1–2 minutes per week per contributor), how AI can merge, deduplicate, and surface trends across those inputs, and how to structure outputs that are actually useful to different audiences — from individual contributors to leadership. I’ll share a concrete example: an interactive system I built where tech leads spend a minute or two summarizing their week, and AI synthesizes those inputs across the team to surface trends that would have taken hours to produce manually.

Attendees will leave with: - A practical framework for capturing organizational learning without creating survey fatigue - Patterns for using LLMs to synthesize distributed signals into coherent, actionable insights - Concrete examples of systems that turn weekly micro-inputs into trend reports automatically - Guidance on designing outputs tailored to different audiences inside your org

This isn’t about the Thoughtworks Radar specifically — it’s about the underlying discipline of knowledge capture and sharing, and how AI makes that discipline accessible to organizations of any size.

I’ve led the Thoughtworks Technology Radar team for the past 18 months, which informs the patterns and principles in this talk. The interactive synthesis system I reference is something I built and have running. Happy to share a demo if that would help the selection committee.

For context: I previously helped organize DevOpsDays Seattle and served on the global DevOpsDays organizing committee, so I have a good sense of what makes a talk land well with this audience.