The Board
For decades, workforce planning worked like this: forecast next year's revenue, divide it by expected output per employee, add a buffer, and submit a headcount request to finance. The plan was built on stable assumptions about what each role produced and what each person cost.
Those assumptions broke in 2024 and have not recovered. When a single AI coding assistant can handle work that previously required two junior developers, and when the regulatory environment might force you to add a compliance function that didn't exist six months ago, an annual headcount number is a fiction. It gives the appearance of planning without the substance.
The shift is already visible in how finance leaders are behaving. Nearly 75% of CFOs are increasing technology budgets in 2026 and 48% are planning double-digit increases. At the same time, only 21% plan staff increases above 4%, down from 31% last year. The money that used to go toward new hires is going toward tools that make existing people more productive, or that replace certain tasks entirely.
The practical response is to move from annual headcount plans to rolling quarterly workforce forecasts, tied directly to revenue projections and product roadmaps. When finance revises the revenue forecast, the workforce model should recalculate automatically. When the product team decides to build an AI feature instead of a manual workflow, the hiring plan should reflect that within weeks, not at next year's planning cycle.
Companies like Mastercard have built internal talent marketplaces that placed 75% of their workforce on a mobility platform, unlocking 100,000 hours of productive capacity and generating $21 million in cost savings. That is what happens when you treat headcount as a pool of capabilities.
Most companies are still faking skill-based planning
The concept is sound and the data backs it up. Skills-based organizations are 63% more likely to achieve high performance than those built around traditional job descriptions. Companies using skills-based assessments report a 91% increase in employee retention. And LinkedIn research shows that a skills-first approach can expand talent pools by up to 15.9 times.
But most companies claiming to do skills-based planning are really doing job-title-based planning with a skills vocabulary layered on top. According to Mercer's 2025/2026 Skills Snapshot Survey, only 38% of organizations maintain an enterprise-wide skills library. Just 55% map skills directly to jobs. And only 21% rate their workforce planning maturity as high.
The gap between ambition and execution is large. 46% of executives cite legacy mindsets and outdated practices as the main obstacle to becoming a skills-based organization. That is a polite way of saying that most managers still frame their hiring requests around job titles. "I need a senior product manager" is how the request lands. "I need someone who can run experiments, interpret quantitative data, and communicate trade-offs to engineers" is what the team actually needs. Those are different searches with different candidate pools.
Making the shift real requires three things. First, build a skills taxonomy that is specific enough to be useful but flexible enough to update quarterly. Starting with an existing framework like ESCO (European Skills, Competences, Qualifications and Occupations) and customizing from there saves months of design work. Second, map every current employee's skills, not just their job title. This is boring, expensive work and there is no shortcut. Third, plan workforce needs as skill gaps. "We need three more people who can do X" is the planning unit that actually produces useful hiring briefs.
Verizon provides an instructive example. The company analyzed 140,000 employees across 11,000 job codes, consolidated them into 10 job families across 2,400 codes, and mapped specific skills to each. That clarity allowed them to build three-year talent forecasting plans that tell them where skill shortages will appear before they become hiring emergencies.
The AI variable makes workforce planning even more important
The practical question for boards is not "will AI replace people?" It will replace some tasks and create demand for others. The WEF's Future of Jobs Report 2025 projects a net gain of 78 million jobs globally by 2030, with 170 million created and 92 million displaced. The question is whether your organization can move people from the displaced category to the created one fast enough. That is a workforce planning problem, not a technology problem.
A practical framework for team composition in 2026 and 2027.
Boards that want to move from reactive headcount management to adaptive workforce planning can apply five principles.
Plan in 90-day cycles, not 12-month cycles. Set a quarterly workforce review tied to financial forecasts and product roadmaps. Treat the annual plan as a directional guide, not a commitment. Recalibrate when inputs change.
Categorize every role by AI exposure. Not "will AI replace this role" but "how will AI change what this role produces in the next 12 months?" A useful three-bucket model: roles where AI amplifies output (the person becomes more productive), roles where AI absorbs tasks (parts of the role disappear), and roles where AI has minimal impact. Budget and plan differently for each bucket.
Invest in reskilling before you need it. The WEF estimates that 59% of the global workforce will need training by 2030, and that 120 million workers are at medium-term risk because they are unlikely to receive it. 85% of employers say they plan to prioritize upskilling, but saying it and budgeting for it are different things. Unilever trained over 23,000 employees in AI skills in 2024 alone. That is the scale of investment that matches the scale of the problem.
Build slack into the system. When you don't know which roles AI will change, carrying a small amount of excess capacity is cheaper than being caught short. Deloitte's scenario planning research recommends building deliberate slack, borrowing from portfolio theory in finance, where diversification absorbs shocks that concentration amplifies.
Treat the skills library as infrastructure. It needs maintenance, updating, and governance, the same way your financial systems do. Mercer's data shows that companies with enterprise-wide skills libraries are progressing steadily (38% adoption, up from 30% in 2023), but the pace needs to accelerate. The companies that have their skills data in order will be the ones that can redeploy people when AI changes the work. The ones that don't will be posting job ads for roles that should have been filled internally.
The companies that plan for uncertainty will outperform those that plan for a future they invented.
Europe's workforce is about to peak. Eurostat projects the EU population will reach 453 million around 2026 before beginning a gradual decline. If participation rates hold, the labour force will shrink by 42.8 million people by 2070. AI is not arriving into an economy with excess labor. It is arriving into one that is running out of it.
That demographic reality makes workforce planning more consequential than it has been in decades. Every person in the organization becomes more valuable, which means deploying them on the wrong work, or failing to develop their skills as the work changes, is a more expensive mistake than it was five years ago.
The 2026 workforce plan is a living model that connects business strategy to skill requirements to hiring, reskilling, and automation decisions, updated quarterly, governed by the same rigor applied to financial planning. The companies that build this capability now will spend the next two years redeploying talent and absorbing change. The ones that don't will spend it posting job ads and wondering why the right candidates never apply.