The Deployment-Transformation Gap
McKinsey’s 2026 HR Monitor, its largest annual benchmarking study surveying 1,300 HR professionals and 5,000 employees across 10 countries, was released this week with a finding that requires immediate attention from the C-suite.
The headline figure sounds like progress: 80% of organisations have deployed AI in at least one HR function. The reality is far less encouraging. Only 20% have actually rebuilt their work processes around the technology.
This is the deployment-transformation gap. Organisations are bolting AI onto existing processes rather than rethinking the processes themselves. They are buying the tools, but they are not changing the work.
This finding triangulates perfectly with the University of Phoenix 2026 C-Suite AI Impact Report, also published this week. That study found that 63% of C-suite leaders have deployed at least one AI use case, but fewer than one-third are transforming workflows around it.
The McKinsey data reveals the operational cost of this gap. Perhaps the most striking finding is the perception divide between HR professionals and employees on learning and development. Nearly one in four workers (24%) report receiving zero formal training in the past year. Meanwhile, HR leaders systematically overestimate both the volume of training employees receive and how much employees value development opportunities .
If HR leaders believe their people are being trained when they are NOT, workforce planning decisions are being made on faulty data. The report confirms this: only 11% of organisations have adopted a long-term workforce planning perspective. The vast majority are planning reactively, cycle to cycle, without a multi-year view of how AI will reshape their talent requirements.
For C-suite leaders, the framework for addressing this is clear:
1) Audit the Gap: Identify where AI has been deployed in your organisation without a corresponding change in workflow. If the process looks exactly the same as it did two years ago, but now includes an AI tool, you have a deployment, not a transformation.
2) Ground-Truth Your Training Data: Do not rely on HR dashboards that may be showing participation rates that do not match employee experience. Conduct direct employee surveys on training access and quality.
3) Extend the Planning Horizon: Move from operational capacity planning to strategic capability planning. The 11% of organisations taking a long-term view of workforce planning have a significant first-mover advantage.
The Infrastructure Cost of Ambition
While the C-suite focuses on workforce transformation, a different kind of constraint is emerging at the infrastructure level.
Gartner published new data this week showing that global data centre electricity consumption is expected to rise by 26% in 2026. AI is driving much of that growth. AI-optimised servers are forecast to account for 31% of data centre electricity use this year, rising to around 37% in 2027, when they are expected to consume more power than conventional servers for the first time.
For leaders, the bigger issue is operational and financial.
AI’s energy demands are becoming a material business risk. As organisations move beyond experimentation and deploy AI across the enterprise, compute requirements rise quickly, along with infrastructure costs, exposure to energy prices, and pressure on already constrained capacity.
CFOs and technology leaders who have not modelled the infrastructure cost of their AI ambitions are working with incomplete numbers. The conversation needs to move beyond “What can this model do?” and ask, “What does it cost to run reliably at scale?”.
That means understanding the unit economics of deployment: compute, energy, infrastructure, integration, maintenance and ongoing usage. Without that visibility, AI budgets can unravel quickly. With it, organisations can make far better decisions about where AI genuinely creates value.
This piece first appeared in Talent Pools Nexus, our weekly read for leaders in research, insights and consulting. Subscribe on LinkedIn.
