Workforce planning has traditionally meant one thing: forecast demand, compare it to supply, and hire (or reorganise) until you have the right people, with the right skills, in the right places, at the right time.
That definition still holds. What’s breaking is the assumption underneath it: that capacity scales linearly with headcount.
Two forces are colliding:
First, skills are shifting faster than classic job architectures can adapt. The World Economic Forum reports employers expect 39% of key skills required in the job market to change by 2030, with technology skills rising fastest (AI and big data, cybersecurity, technological literacy), alongside human capabilities like creative thinking and resilience.
Second, automation is moving deeper into knowledge work. McKinsey Global Institute argues current generative AI capabilities (combined with other technologies) could automate a substantial share of work activities and lift productivity growth—if organisations redesign work and redeploy time effectively.
The implication is uncomfortable but liberating: future capacity planning is as much about work design (what gets automated, orchestrated, or augmented) as it is about recruiting.
“Agentic AI” changes the planning question—but hype can destroy budgets
You’ll hear new language in 2026: “AI agents”, “agentic workflows”, “digital workers”. The underlying idea is consistent: systems can increasingly execute multi-step tasks with less human prompting. HR and TA are obvious early targets—screening, scheduling, candidate communication, skills inference.
But the market is noisy. Gartner has warned that over 40% of agentic AI projects may be cancelled by 2027 due to escalating costs, unclear value, or inadequate risk controls—and it explicitly calls out “agent washing” (rebranding assistants, RPA, or chatbots as agents without real capability).
So the workforce-planning question is not “How do we buy AI?” It’s:
Where does digital execution create measurable capacity—and where does it create new risk?
A practical framework: plan for a blended workforce, not a binary one
A useful way to think about modern workforce planning is a portfolio model:
Human judgement work (high-stakes decisions, relationship building, negotiation, ethical accountability). This is where you should concentrate your best people.
Human-with-AI work (analysis, drafting, synthesis, scenario modelling, skills matching—humans remain responsible, AI accelerates). This is where AI literacy and workflow design generate productivity.
AI execution work (repetitive, rules-based, high-volume steps with clear guardrails and auditability). This is where throughput explodes—if governance exists.
This isn’t just theory. Organisation for Economic Co-operation and Development surveys across multiple countries found four in five AI users reported improved performance at work, but also identified worker concerns: job loss, wage pressure, work intensity, and data collection—highlighting why training and consultation correlate with better outcomes.
Regulation is becoming a planning input
If your workforce plan assumes “we’ll automate screening and performance decisions”, you’re no longer only making an operational choice—you’re creating regulatory obligations.
The European Commission’s AI Act timeline is explicit: the Act entered into force in August 2024; prohibited AI practices and AI literacy obligations have applied since February 2025; obligations for general-purpose AI models became applicable in August 2025; and the Act becomes fully applicable in August 2026, with some extended transitions for specific categories.
In employment contexts, prohibited practices include emotion inference in workplaces (with narrow exceptions).
In addition, employment-related AI may fall into “high-risk” categories, triggering obligations around documentation, oversight, and monitoring, depending on the system’s function and impact.
For workforce planning, that means your “tech roadmap” and your “people plan” now share the same governance layer.
What a CPO/CFO-ready workforce plan looks like in 2026
The most credible workforce plans we see (especially in AI/IT-heavy organisations) have five components:
A skills inventory that’s usable, not aspirational. Not a spreadsheet of buzzwords—something that can answer: Which critical capabilities do we have, at what depth, and where are the gaps? This aligns with WEF’s emphasis on mapping skills and capability needs as a pillar of workforce transformation.
A “work decomposition” view. Break key functions into activities and decide which are candidates for automation vs augmentation vs human ownership. McKinsey’s analysis is useful here because it frames automation at the activity level, not job title level.
A technology plan that includes process redesign. Workforce transformation can’t be treated as a standalone tech upgrade; it must align vision, skills, technology, process, and culture.
A risk model with named owners. Use a recognised risk framework language—e.g., National Institute of Standards and Technology defines an AI system as a machine-based system that generates outputs such as predictions or recommendations and notes AI systems operate with varying levels of autonomy; it frames risk management as essential to trustworthy AI.
A talent strategy that recognises the new scarcity. If two-thirds of employers plan to hire talent with specific AI skills, competition is structural, not cyclical.
That shifts workforce planning into a continuous discipline: build, buy, borrow (contract), and redeploy—often at the same time.
For Acquism clients, this is where recruitment becomes a workforce strategy lever, not a transactional service. If you’re building AI engineering, MLOps, data platforms, or applied ML teams, you need a plan that can answer two questions at once: “What skills do we need?” and “What work can we redesign so humans spend time where humans are uniquely valuable?”