Why Jack Dorsey’s 40% AI layoffs are a wake-up call for HR

Block has been rolling out its own AI agents internally, and the company says engineers are already saving material amounts of time with these tools. That kind of concrete productivity story is what CEOs will now demand from HR and learning functions. Generic “AI literacy” programs won’t be enough.

Dorsey has spoken… what next?

First, conduct an AI-exposure audit by role and process. HR should partner with technology and operations leaders to build a clear catalog of AI and automation already in use. Then, at the team level, leaders should identify which tasks are already being partly automated, which could be, and what that means for staffing over the next 12 to 24 months.

Second, put governance around AI and workforce decisions. An “AI and workforce” council that brings together HR, legal, risk, technology and business leaders can evaluate the people impact of major AI deployments before they scale. Such a group can set principles for when automation should lead to redeployment versus reduction, and can ensure that the company doesn’t take on undue operational or reputational risk in pursuit of rapid efficiency gains.

Finally, take control of the narrative. When a CEO publicly argues that “intelligence tools” allow “a significantly smaller team” to do more, employees, candidates and customers will fill in the blanks. Clear communication is essential about where AI is likely to change work, where human roles remain central, what the time horizon looks like, and how the company will treat people if restructuring becomes necessary.

The Block story stands out because it is a clear signal that AI will increasingly be used as a rationale for doing more with fewer people. Dorsey’s claim that a smaller, AI-enabled staff can outperform a larger one will be tested in the quarters ahead, and the outcome will vary from company to company and industry to industry. For now, executives have a narrow window to prepare their organizations for a potential future in which leaner, AI-native operating models become the norm rather than the exception.