Panel Discussion: The Agentic Software Factory That Enterprises Actually Need

This panel discussion stays refreshingly honest from start to finish. 𝐇𝐞𝐞𝐧𝐚 𝐏𝐮𝐫𝐨𝐡𝐢𝐭, Director of AI Startups at Microsoft, 𝐂𝐡𝐚𝐬𝐞 𝐍𝐨𝐫𝐭𝐨𝐧, Head of AI at Faros AI, 𝐈𝐭𝐚𝐦𝐚𝐫 𝐅𝐫𝐢𝐞𝐝𝐦𝐚𝐧, CEO and Co-founder at Qodo AI, and 𝐏𝐚𝐭𝐫𝐢𝐜𝐤 𝐕𝐮𝐨𝐧𝐠, Director of Product Management at Moderne, examine what separates a demo-ready coding agent from an agentic software factory an enterprise can actually run. The first wave of developer AI helped individuals write code faster; the next wave changes the entire delivery lifecycle. Learn why generating more code faster is the wrong goal, and what quality, control, reliability and accountability look like at scale. The panel also digs into the work enterprises cannot skip - modernizing large legacy codebases, holding standards across thousands of repositories, validating AI-generated changes, handling security and compliance, resolving production issues, and measuring whether more AI usage is actually producing better engineering and business outcomes. As they put it, the goal is not tokenmaxxing, it is outcome-maxxing. 𝐇𝐢𝐠𝐡𝐥𝐢𝐠𝐡𝐭𝐬: 0:00 Would 10x More Code Produce 10x More Value? Probably Not 1:05 The Real Enterprise Question: Can You Trust AI-Generated Code? 2:38 Panel Intros - Heena Purohit, Chase Norton, Itamar Friedman, and Patrick Vuong 5:13 What Broke First as AI Code Scaled Across Organizations 6:44 The Bottleneck Moved From Writing Code to Trusting Code 8:16 What AI SDLC / Agentic SDLC Actually Means 9:18 Judgment Spilled Downstream Onto Senior Developers 10:19 Enterprise Reality: Thousands of Repos, Legacy Code, CI/CD, Compliance 11:52 The 'Dark Factory' Vision and Working Backwards From It 14:27 Code as Data: Why Agents Need a Compiler-Accurate View of the Codebase 15:59 Context Is King - Company Context Lakes, Guardrails, Verification Spaces 18:32 Bringing Determinism and Control Into a Probabilistic Stack 21:08 Observe Before You Optimize, Then Govern 23:39 Breaking the SDLC Into Small Components to Reduce Variance 25:13 What Does Quality Even Mean? Breaking It Into Measurable Policies 27:19 From Token Maxing to Outcome Maxing - What to Actually Measure 30:52 Measure What Lands, Not What the Factory Consumes 32:26 The ROI Equation: Features and Infra Over Incidents, Bugs and Cost 35:00 Where Should Agents Be Autonomous - and Where Not? 38:04 A Review Panel of Agents Encoding Senior Engineers' Tribal Knowledge 41:08 Greenfield vs Brownfield and Building High-Confidence Verification Loops Register for TestMuConf 2027: https://www.testmuai.com/testmuconf-2027/?utm_source=youtube&utm_medium=organic&utm_term=&utm_campaign=agentic_software_factory_enterprises_actually_need #TestMuConf #TestMuAI #AgenticEngineering #EnterpriseAI #SoftwareFactory #DevOps #AIAdoption #Panel