Building AI Agents with Real ROI in the Enterprise SDLC
AI Engineer Summit 2025 · 20:56
← All organizationssourcegraph.com ↗
Code intelligence and enterprise developer tools
Sourcegraph
Sourcegraph builds code intelligence tools that help developers search, understand and change large codebases across repositories. Code Search supports regular expressions and searches across commits and diffs, while Deep Search explores code in response to natural-language questions and returns sourced answers. Code navigation lets developers follow definitions and references; Batch Changes automates changes across repositories, and Code Insights tracks migrations, version usage and vulnerability remediation.
Launched in 2013, Sourcegraph developed around universal code search. In December 2025, its AI coding product Amp announced a spinout into an independent research lab, Amp Frontier Corporation. Alongside that announcement, Sourcegraph’s vice president of business, Dan Adler, moved into the CEO role, while Sourcegraph co-founder Quinn Slack became Amp’s CEO. Sourcegraph’s stated focus under the announced spinout plan was code search and code understanding for large codebases.
Sourcegraph Enterprise is the company’s primary offering, with cloud and self-hosted deployment options. Integrations bring search and navigation into editors, code hosts and review tools, while its Model Context Protocol server connects AI tools to the same code intelligence capabilities.
4 talks
Newest first
The AI emperor has no DAUs: why most devs still don't use code AI
AI Engineer World's Fair 2024 · 18:45
3 speakers at AIE
Affiliations reflect their AIE appearances, not necessarily current employment.
Start here
- The Code AI Maturity Model and What It Means For You
Start here for vocabulary to distinguish assistance with individual coding tasks from automation of broader development workflows.
Ado KukicAI Engineer Summit 2023
- The AI emperor has no DAUs: why most devs still don't use code AI
Learn how survey scope, paid subscriptions, and active-user measures can produce different pictures of coding-tool adoption.
Quinn SlackAI Engineer World's Fair 2024
- The ROI of AI: Why You Need Eval Frameworks
Use this talk to examine productivity measurement beyond intuition, including enterprise examples and lessons from The Mythical Man-Month.
Beyang LiuAI Engineer World's Fair 2024
- Building AI Agents with Real ROI in the Enterprise SDLC
Booking.com's Bruno Passos and Sourcegraph co-founder Beyang Liu explain how their teams apply context-aware coding assistance and software-development agents to enterprise codebases while measuring business impact.
Bruno Passos · Beyang LiuAI Engineer Summit 2025
Messages from the stage
A framework for increasing autonomy
Ado Kukic’s maturity model distinguishes autocomplete, codebase-aware assistance, requirements-driven feature generation, and progressively more autonomous development. Cody examples illustrate the framework rather than establish that every level has been achieved.
Verifiable value and sustained use
Quinn Slack questions whether surveys and subscription figures capture habitual use. His discussion connects developer hesitation and the limits of ghost text to product design that makes assistance easy to verify.
Evaluating assistance in enterprise codebases
Beyang Liu discusses code search, relevant context, and engineering KPIs as inputs to evaluating coding assistants. The joint session with Bruno Passos extends the discussion to modernization metrics, feature-flag cleanup, and automated code review.
Affiliations reflect each recorded session, not necessarily current employment. The enterprise SDLC case study is a joint discussion with Booking.com's Bruno Passos.