Boards keep asking the wrong question. Not "will AI take my people's jobs?" but "which specific tasks does AI already perform better than a human – and what does that mean for us?". These are not the same question. And the difference between them determines whether an AI deployment ends in success or yet another costly pilot with no results.
AI Replaces Tasks, Not Positions
Gartner estimates that every year more than 30 million job roles are being redesigned – not eliminated – by AI. This is a crucial distinction that industry headlines consistently miss. A single role can simultaneously lose 20–40% of its existing tasks and gain new ones: coordination, AI oversight, decision-making with incomplete data. In other words – the job description, KPIs and performance review system become outdated before HR even notices.
From my observations – confirmed by data in the Deloitte "Navigating the end of jobs" report – the companies that achieve real value from AI do one thing: they break roles down into tasks, and only then decide which are performed by a human, which by AI, and which by a human + machine collaboration model. This is not digital transformation. It is a redesign of how work itself is defined.
The Biggest Risk Isn't Technology – It's the Skills Gap
Technology becomes available faster than organizations can absorb it. Gartner forecasts that by 2030, 75% of work will be performed by AI-augmented humans, and 25% by AI autonomously. Until then, however, organizations will be grappling with something entirely different: decision-making chaos, "shadow AI" – employees using AI outside IT's control – and overestimating short-term gains while underestimating transition costs.
Deloitte points out that organizations investing in systematic AI upskilling are 2.5 times more likely to achieve positive business outcomes. Not because of the technology – but because the people in those companies know how to work with it.
What Boards Should Do – Now
Don't ask: "How do we replace people with AI?" Ask instead: "Which tasks in our organization can AI take over today – and how do we redesign roles around what remains uniquely human?"
Three steps that make sense regardless of industry:
- Break roles down into tasks and assess their susceptibility to automation.
- Build AI literacy not just in IT, but across the entire organization.
- Change incentive systems to reward outcomes, not manual effort.
Because AI won't eliminate work. It will eliminate organizations that never learned to work differently.
Dear Reader, if you believe the topic described above applies to your company and you'd like to discuss how to effectively redesign roles and competencies in the AI era with your Board – I invite you to get in touch. Leszek Giza.