You Added AI Agents. Why Are You Still Running a Separate Quality Organization Like It Is 2009?
If you are adopting AI and still defending a separate quality organization in 2026, you are not making a technical argument. You are making an emotional one. CEOs and boards should treat that as a leadership warning.
Operating models must evolve to align with new capabilities, especially in an agent-driven development environment.
Modernize the Quality Operating Model
Quality shifts from a downstream gate to an inherent property of the development system itself. This requires embedding quality expertise within engineering teams rather than maintaining it as a separate organizational unit.
Agents require immediate feedback loops to learn and iterate effectively. Inter-departmental handoffs for quality assurance introduce latency, undermining the core advantage of agentic workflows.
Confidence in software quality now derives from continuous instrumentation, automated validation, and system-generated proofs, replacing reliance on manual inspection and ceremonial sign-offs.
The return on investment for separate quality organizations diminishes as automated testing, observability, and embedded quality practices become integral to the engineering process.
The first question for any AI program: does the organizational structure support rapid feedback and continuous validation, or does it enforce antiquated gates?
Operating models must evolve to align with new capabilities, especially in an agent-driven development environment.
Modernize the Quality Operating Model
Quality shifts from a downstream gate to an inherent property of the development system itself. This requires embedding quality expertise within engineering teams rather than maintaining it as a separate organizational unit.
Agents require immediate feedback loops to learn and iterate effectively. Inter-departmental handoffs for quality assurance introduce latency, undermining the core advantage of agentic workflows.
Confidence in software quality now derives from continuous instrumentation, automated validation, and system-generated proofs, replacing reliance on manual inspection and ceremonial sign-offs.
The return on investment for separate quality organizations diminishes as automated testing, observability, and embedded quality practices become integral to the engineering process.
The first question for any AI program: does the organizational structure support rapid feedback and continuous validation, or does it enforce antiquated gates?
After 20 years in software development, Norman is both a hands-on leader and defining the new age of AI SDLC for some of the biggest brands in the world — and exploring it with the builders. He writes here about things he is hearing and seeing. All posts are his personal points of view and do not reflect any employer or any customer he has ever had contact with.
The views and opinions expressed in this article are the author’s own and do not represent the positions of any employer, client, or affiliated organization.
