Redesigning How Software Gets Built With AI Agents — Sonar & McKinsey Panel

Most companies say they've deployed AI agents. Fewer than a third say it's paying off. What do the rest do differently? A panel with Sonar CEO Tariq Shaukat and Ali-Reza Adl-Tabatabai, former CEO of Gitar (now part of Sonar) and former head of Uber's developer platform, moderated by McKinsey. Research with Sonar found the teams getting real results from agents change three things: their process, their tooling and, most often overlooked, their people and roles. Tariq explains why lighthouse teams alone don't spread new habits, and why engineers need to make implicit knowledge explicit and get great at critical thinking, because AI always sounds sure. Ali covers how big companies can borrow from small teams through their platform orgs, and walks through how trust builds step by step: from AI reviews, to blocking on issues, to autofix, to auto-approve and auto-merge. In this talk: • Three factors behind teams that scale agents: process, tooling and people • Why role models and lighthouse teams aren't enough on their own • The skills that matter now: explicit thinking, critical thinking, generalists • How trust grows from AI code review to fully automated merges SPEAKERS Tariq Shaukat, CEO, Sonar X: https://x.com/tariqshaukat Ali-Reza Adl-Tabatabai, Former CEO & co-founder, Gitar (now part of Sonar) LinkedIn: https://www.linkedin.com/in/aliadl/ X: https://x.com/aadltaba LINKS Sonar: https://www.sonarsource.com Gitar by Sonar: https://www.sonarsource.com/products/gitar/ Sonar on X: https://x.com/SonarSource CHAPTERS 0:00 Intro 0:57 Research: what helps teams scale 1:51 Process, tooling and people 3:04 Meet the panel 3:19 Lessons from Sonar's own transition 5:38 Lighthouse teams aren't enough 6:18 What big companies can learn from small teams 8:37 Measure first 9:36 Automate whole workflows 10:21 Automate the grunt work 11:04 What this means for people and orgs 12:43 Make the implicit explicit 13:08 Critical thinking 14:28 The era of the generalist 15:07 Will we trust agents to validate code? 16:02 From reviews to blocking to autofix 17:52 Auto-approve and auto-merge 18:36 Where this is heading Recorded at the AI Engineer World's Fair 2026 in San Francisco. Subscribe for more talks from the engineers building with AI. AI Engineer: https://ai.engineer YouTube: https://www.youtube.com/@aiDotEngineer X: https://x.com/aiDotEngineer LinkedIn: https://www.linkedin.com/company/aidotengineer/ #AIAgents #SoftwareEngineering #AIEngineer