The great AI transformation of 2025 isn’t a technology challenge—it’s a leadership challenge. The tools are already here. What’s missing is the executive vision and operational discipline to use them. Shopify CEO Tobi Lütke has taken a firm stance: no new hires unless you can prove AI can’t do the job.
In a memo shared internally and posted publicly, Lütke announced that headcount requests at Shopify must be preceded by an evaluation of automation. “Prove that AI can’t do it” is now a condition for team growth. AI proficiency will be included in performance reviews. Managers are expected to explore automation before seeking additional human resources. At Shopify, AI is no longer a support tool—it’s the default teammate.
This is not a minor operational tweak; it reframes headcount planning, performance management, and workflow design around a single question: can software replace this job function? Back in 2023, Shopify launched “Sidekick,” an AI assistant for merchants that creates discount codes, generates reports, and automates routine tasks. Internally, Lütke wants the same standard to apply; teams are expected to redesign their work assuming autonomous agents are built in from the start.
Lütke says this is the fastest shift in work practices he’s seen in his career, yet the policy raises a cascade of questions. Who decides when AI has failed? What metrics determine when a task truly requires a human? Are we building efficiency—or deferring hard decisions to systems that may not be flexible enough (or ready) to do the job?
We may be witnessing the emergence of a new leadership doctrine: Assume AI. Prove the need for humans.
Author’s note: This is not a sponsored post. I am the author of this article and it expresses my own opinions. I am not, nor is my company, receiving compensation for it. This work was created with the assistance of various generative AI models.