AI is changing the talent game by changing work itself: which tasks people do, which skills employers value, and how organizations develop and deploy their people. The response is not simply to hire AI specialists or cut jobs. Employers are combining work redesign, training, selective hiring, redeployment and new rules for human oversight. For workers, that shift can affect career paths, pay and job security; for leaders, it makes workforce planning part of the AI investment decision.
AI is changing tasks before it changes whole jobs
A job title can conceal very different effects. In one role, AI may draft routine material or summarize information; in another, it may help an employee analyze more options or respond faster. Some tasks are automated, some are augmented, and many jobs are being redesigned around review, judgment, exceptions, relationships and accountability. New responsibilities are also emerging in areas such as model evaluation, AI governance, data stewardship, security and workflow supervision.
As an Amazon Associate I earn from qualifying purchases.
This is why leaders need to analyze work at the task level rather than assume an entire occupation is either safe or replaceable. The same organization can automate certain activities, hire specialists, retrain current employees and reduce some roles at the same time. The World Economic Forum reports that 86% of surveyed employers expect AI and information-processing technologies to transform their business by 2030; that is an expectation, not a measured outcome. In the same survey, 41% said they expect to reduce their workforce where AI can replicate roles, while 77% plan to reskill or upskill existing workers. The WEF figures describe employer plans and expectations, not proof of how many jobs will ultimately be lost or training programs completed.
Four kinds of capability employers need
- Technical AI skills: machine learning, model development and evaluation, data engineering and cybersecurity.
- Applied AI skills: choosing tools, designing workflows, supervising agents, checking outputs and automating processes safely.
- Domain expertise: understanding the business, customers, systems and rules in fields such as finance, law, medicine, sales or manufacturing.
- Human and managerial skills: communication, judgment, collaboration, coaching, negotiation, ethical reasoning and change leadership.
AI fluency is not a substitute for the other three. A person who can use a tool but cannot assess whether its answer is sound may create risk rather than value.
#1 Best Overall
AI use is moving faster than workforce preparation
The gap between using AI and being prepared to use it well is a central workforce challenge. The Conference Board reported that 55.1% of surveyed workers use generative AI or AI agents daily or weekly, while 33.3% had taken employer-provided AI training in the previous six months and 28.3% said their organization offered no AI training. These are worker-reported survey findings, not a count of all workers worldwide. The same report presents separate measures for access to time, tools and resources; it does not establish one combined readiness score. The Conference Board’s findings point to a practical problem: employees may be expected to use AI before employers have supplied training, access, time or clear standards.
Adoption and value are also different things. In PwC’s 29th Global CEO Survey, fewer than one-quarter of CEOs reported extensive AI use across major business areas, and 22% said their business was highly exposed to a lack of key skills. “Extensive” is the survey’s wording; the result should not be read as broad, mature deployment across companies. PwC’s workforce analysis underlines why tool access alone is not a workforce strategy.
Leaders are combining redesign, development and selective hiring
The most credible talent response is a portfolio: build skills internally, buy capabilities that are scarce or urgent, borrow expertise where temporary speed matters, and redeploy employees as tasks and roles change. The World Economic Forum reports that surveyed employers plan both to reskill existing workers and recruit for AI-related capabilities by 2030: 69% expect to recruit people skilled in designing or improving AI tools, and 62% expect to hire people with skills for working with AI. Those plans can complement, rather than contradict, employee development.
Rank #2
- we like to ship out right away
Redesign workflows before making headcount decisions
Start by asking which tasks AI should perform, which should remain with people, and where human review is required. Map a workflow from end to end, identify repetitive or information-heavy tasks that can be checked, and distinguish them from work requiring contextual judgment, accountability, empathy, negotiation or physical presence. Then adjust roles, decision rights, performance measures and manager responsibilities to match the new division of labor.
AI value is often created across a workflow, not by adding an isolated tool to one department. KPMG’s AI Quarterly Pulse Survey frames transformation at that broader level. If an organization changes a task but leaves approvals, incentives and handoffs untouched, the new capability may simply shift bottlenecks elsewhere.
Build skills and make them visible
Skills-based workforce planning asks what capabilities the company has, what its changing workflows need, which employees can be trained, and where outside hiring is necessary. Mercer’s 2025/2026 Skills Snapshot Survey found that 91% of companies surveyed see AI transforming their workforce, 38% maintain a single enterprise-wide skills library and 55% map skills directly to jobs. The figures are evidence of movement, not proof that every organization has a reliable skills system. Mercer’s survey suggests that skills architecture is becoming workforce-planning infrastructure, rather than just a feature of HR software.
Rank #3
A skills library is useful only if it stays connected to real work. Leaders need to refresh it as workflows change and use skills information in internal hiring, learning plans, promotion and career development. A taxonomy that is out of date can make employees seem unqualified for work they could do with modest training.
Upskill, reskill and redeploy for different purposes
- Upskilling adds capabilities to help someone do their current role differently or more effectively.
- Reskilling prepares someone for substantially different work.
- Redeployment moves an employee into another role or workflow where their experience is useful.
- Career-path redesign creates new progression routes when traditional junior tasks change or disappear.
Existing employees may bring customer knowledge, institutional context and process experience that a new specialist needs time to acquire. At the same time, some capabilities—such as foundational AI engineering, data infrastructure, security or governance—may be too scarce or time-consuming to develop internally. Mercer reports that 65% of executives expect 11%–30% of their workforce to be redeployed or reskilled because of AI over the following two years. That is an executive expectation reported by Mercer, not a measured forecast of actual moves. Mercer’s 2026 Global Talent Trends report also describes organizational redesign plans, which should likewise be understood as plans rather than completed changes.
Hire for demonstrated hybrid capability
Hiring criteria should reflect the work to be done, not a list of fashionable tool names. A strong candidate may combine domain knowledge, practical AI fluency, the ability to verify and challenge outputs, communication skills, comfort with ambiguity and an understanding of privacy and security. Work samples or workflow demonstrations can reveal more than résumé keywords.
KPMG reports that 65% of organizations in its survey are investing in upskilling and reskilling, that 54% say social and interpersonal skills are more important than purely technical ones, and that strong AI talent commands a reported 6%–15% salary premium. These are KPMG survey signals, not universal measures of employer behavior, skill value or compensation. The KPMG survey supports a blended talent model, not the conclusion that human skills replace technical expertise.
Human skills and career development matter more, not less
AI can generate, summarize or analyze, but people still have to decide what matters, explain choices, coordinate with colleagues and take responsibility for consequential decisions. Communication, adaptability, problem-solving and judgment become especially important when outputs are fast but uncertain. A GMAC employer survey similarly emphasizes the continuing importance of human skills alongside AI. GMAC’s findings are best read as a case for combining capabilities, not as evidence that technical skills have become unimportant.
There is a career-development risk hidden in efficiency gains. Entry-level employees often learn through drafting, research, basic analysis, documentation, customer triage and administrative coordination. If AI takes over much of that work, organizations may save time now while reducing the practice that produces future experts and managers. Some junior tasks are routine; others are developmental because they teach how the business works and how to spot errors.
Best Value
- Author: Bungay Stanier, Michael.
- Publisher: Page Two
- Pages: 244
- Publication Date: 2016-02-29
- Edition: 1
A survey of 546 U.S. HR leaders conducted by Morning Consult for D2L in January 2026 points to structured learning, internal apprenticeships, rotations, training simulations and skills-based hiring as ways to adapt entry-level pathways. The D2L survey does not establish that traditional entry-level jobs have disappeared. It does raise a leadership question: if AI performs more junior work, where will early-career employees get supervised practice, feedback and exposure to foundational tasks?
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Trust and expectations are part of the implementation
Employees need to know why AI is being introduced, which tasks or roles may change, how performance expectations will shift, what support is available and what data about their work will be monitored. They also need a way to question or appeal decisions that rely on AI. Clear communication does not eliminate uncertainty, but silence can leave people to assume the worst.
Mercer reports that 40% of employees in its 2026 data were concerned about job loss due to AI, compared with 28% in 2024; 62% said leaders underestimate AI’s emotional impact, while 19% of HR leaders said those effects were part of their digital implementation strategy. These are Mercer survey responses, not a census of employee sentiment. KPMG separately reports that employee resistance in its survey increased from 5% to 20% between quarters, with trust and ethical concerns cited as a major driver. Mercer’s survey and KPMG’s survey make the same operational point: training alone will not resolve concern about job security, surveillance, fairness or accountability.
A practical 90-day plan for leaders
Days 1–30: Diagnose the work
- Choose a small number of high-value workflows tied to business needs, rather than starting with a tool rollout.
- Break affected roles into tasks and identify where AI might automate, augment or create new work.
- Map critical skills, current employee capabilities and hard-to-fill gaps.
- Audit existing AI use, including informal or unapproved tools, and identify data, privacy, security and quality risks.
- Identify entry-level tasks that provide important learning and decide how that experience can be preserved.
Days 31–60: Design the workforce response
- Define the future workflow, including which decisions require human review and who is accountable for the result.
- Create role-specific learning paths with practice on real work, not just introductory tool instruction.
- Decide which capabilities can be built internally, which require selective hiring and which can be borrowed temporarily.
- Set acceptable-use, data-handling, security and quality-control rules that employees can apply in day-to-day work.
- Clarify manager responsibilities, performance expectations and routes for internal mobility.
Days 61–90: Pilot, learn and measure
- Pilot redesigned workflows with employees who will use them and give them protected time to learn.
- Track cycle time and productivity together with accuracy, rework, customer outcomes and employee confidence.
- Measure whether employees apply new skills and whether the pilot improves proficiency or internal movement.
- Collect worker feedback, investigate failures and adjust the workflow, training or rules before expanding.
- Scale only when the workflow works in practice and the organization can support it safely.
How to tell a real talent strategy from AI theater
Usage counts alone—logins, prompts or tokens—show activity, not whether work improved. Leaders should measure outcomes alongside skill development and workforce health, and compare results against a clear baseline where one is available.
- Business outcomes: productivity, cycle time, quality, error rates and customer results for the redesigned workflow.
- Workforce outcomes: time to proficiency, skills gained and applied, internal fill rates, retention of critical talent and successful redeployment.
- Career outcomes: progression from entry-level roles, access to supervised practice and the strength of future talent pipelines.
- Employee experience: access to training and tools, confidence, trust and clarity about changed expectations.
- Controls: clear accountability, privacy and security safeguards, output verification and a process for raising concerns.
A weak program tells people to use AI without approved access, training time, workflow changes or a way to verify outputs. A strong one connects a business need to redesigned work, gives employees the capabilities and support to do it, and checks whether the result benefits customers and the organization without quietly eroding future expertise.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
