As artificial intelligence (AI) becomes part of everyday work, organisations face a question that goes beyond choosing technology: do their people have the skills to use it well? Access to an AI tool does not automatically mean better work. Employees need to know when to use it, how to judge its output, what information must stay protected and where human judgement is essential.
That makes reskilling a business-planning issue, not just a training-calendar activity. Companies that wait until a skill shortage becomes urgent may find themselves competing for scarce talent while their own employees are unsure how their roles are changing.
PeopleStrong’s AI Ground Report 2026, a survey of more than 500 HR leaders across 16 industries, found that 74% see AI-led HR evolution as a priority and 84% have begun some form of AI activity, yet fewer than half of active implementers are satisfied with the results. Foundations and integration matter as much as adoption.
Start with tasks, not job titles
A job title rarely tells the full story of what a person does. Break important roles into tasks and ask three questions: which tasks could be automated, which could be augmented by technology, and which still depend heavily on human judgement, relationships or physical work?
An AI tool may help a recruiter draft outreach messages or summarise interview notes, but it does not remove the need to understand a role, ask the right questions, weigh evidence, protect candidate information and build trust. AI may help a supervisor analyse production data, but it does not replace reading site conditions and making safe decisions.
This task-level view focuses learning investment on the capabilities that are actually changing.
The skills that matter most
Priorities vary by function, but most organisations should consider a combination of:
- AI and digital fluency: understanding what approved tools can do, how to give them context and how to verify outputs rather than accepting them automatically.
- Data literacy: reading dashboards, understanding basic trends, questioning data quality and separating correlation from cause.
- Critical thinking and judgement: spotting gaps, weighing alternatives, recognising risk and knowing when to escalate.
- Communication and collaboration: explaining decisions, working across teams and handling situations that need empathy and clarity.
- Role-specific expertise: the technical, operational, commercial or regulatory knowledge needed to apply new tools safely in a particular job.
These capabilities work together. AI fluency without subject knowledge can produce confident but unsuitable answers; subject knowledge without willingness to adapt limits the value of new tools.
Build a reskilling programme in five steps
1. Map critical roles and skill gaps
Identify the roles most affected by business change, the skills that will matter more and the capabilities already available internally. Prioritise a manageable number of role families rather than retraining everyone on every new technology.
2. Build learning around real work
Combine short learning modules with supervised application, job aids, team demonstrations and time to experiment in approved environments. People learn faster when they apply a new capability to a relevant task and receive feedback.
3. Equip managers to support adoption
Managers shape whether people feel able to ask questions, test new methods and admit mistakes. Give them guidance on setting expectations, reviewing AI-assisted work and recognising when more training or human intervention is needed.
4. Create opportunities to use the skill
Training loses value when employees return to roles where the new capability is neither expected nor supported. Identify projects, redesigned tasks, internal moves or stretch assignments where people can apply what they have learned.
5. Measure capability, not just attendance
Course completion shows participation, not proficiency. Use demonstrations, quality checks, manager feedback and business outcomes to see whether a skill is being applied, and update the learning plan as tools and tasks change.
Make reskilling part of workforce planning
Link learning priorities to hiring plans, technology roadmaps and future workload. Some gaps can be closed by training existing employees; others need external hiring, specialist support or a redesigned role. A realistic plan combines these approaches rather than treating recruitment and learning as separate activities.
Be open with employees about why work is changing, what skills are needed and what support is available. Avoid promising that every role will stay the same, but equally avoid assuming that exposure to technology means a role will disappear. The impact depends on the tasks, the business’s choices and how the technology is deployed.
Frequently asked questions
What is the difference between reskilling and upskilling?
Upskilling deepens skills for a person’s current role; reskilling prepares them for a different or substantially changed role. Most AI programmes need both.
Which roles should we reskill first?
Those where tasks are changing fastest and the business impact is highest. A task-level review of critical role families shows where to start.
How do we know whether reskilling is working?
Look for applied capability: work samples, quality checks, manager feedback and business outcomes, rather than course completion alone.
Prepare people to work differently
The future-ready workforce is not the one that uses the most AI. It is the one that combines technology with domain knowledge, judgement, ethical conduct and continuous learning. Organisations that build those capabilities deliberately will adapt faster as work changes.
Source: PeopleStrong, AI Ground Report 2026 (June 2026).