DORA: AI amplifies teams — so grow people who own AI output

The piece is the final installment of a three-part series. The first part argued that AI at work supplements effort rather than ability; the second argued that what people should learn to keep their judgment is the unchanging foundation, not AI trends. This third part picks up where individual judgment runs out and asks how organizations can cover the rest.

The author's central distinction is between roles that merely stand between steps without judgment, which AI can replace, and people who can judge AI output and take responsibility, who are not made unnecessary. Responsibility for delivered work rests with the organization, not the AI, so the organization needs people who can judge within their own specialty and combine specialties to cover gaps. Small organizations that cannot keep every expert in-house can at least decide in advance which field is checked by whom.

On what organizations can do, the author draws on the DORA report's view that investing in organizational systems matters more than installing tools, and proposes three steps: building flows that keep human responsibility for AI output, fostering a culture that values judgment and explanation over speed, and investing in developing people. The warning on the third point is that cutting headcount and handing junior work to AI may work now but will leave no one able to judge as AI advances.