Skills-Gap Audit: Workhuman's Guide to Reskilling for AI

A skills-gap audit is the first step in reskilling for AI. Follow Workhuman's plan: on-the-job training, peer learning and paired upskilling.

By Cosby Khosa · Published 21 October 2026

Key takeaways

Start with a skills-gap audit that compares future skill needs with current skills.
Use on-the-job training and peer learning to build skills at low cost.
Pair upskilling with reskilling so you can strengthen current roles and prepare people for new ones.
Reskilling frameworks are hard to scale, and AI-driven skills forecasting is an emerging tool to be used with human judgement.

A skills-gap audit is the first step in Workhuman’s guide to reskilling for AI. The how-to piece, Reskilling in the Age of AI: How to Future-Proof Your Workforce, was published around 14 October 2026. It proposes the audit first, then on-the-job training and peer learning, and a pairing of upskilling with reskilling.

It also lists a challenge, which is building reskilling frameworks that scale, and mentions AI-driven skills forecasting as an emerging tool.

This article turns those points into a practical plan that HR leaders and managers can follow. Our own examples are marked as such. Workhuman sells employee recognition and engagement software, so its advice also serves its business. We discuss that near the end. The article is written for employers, and for workers who want to understand what their employer might offer.

Why a skills-gap audit comes first

Before you train anyone, you need to know what skills you have and what skills you need. A skills-gap audit does exactly that. It compares the skills your organisation will need in the near future with the skills it has now.

Starting here saves money and time. Without an audit, training can be scattershot. People may take courses that do not match business needs, while real gaps stay open. Money is wasted, and staff become cynical.

It builds a shared picture. When leaders, managers and employees see the same data, they can discuss priorities and plans. It reduces guesswork and arguments.
It supports fairness. If decisions about training and moves are based on clear skills data, they are less likely to rest on favouritism or assumptions.
It reveals hidden strengths. Employees may have skills that are not recorded in their job titles, such as languages, coding or project management learned elsewhere. These can be used.

Workhuman’s emphasis on starting with a skills-gap audit reflects good practice in workforce planning. The next section explains how to run one.

How to run a skills-gap audit step by step

A skills-gap audit does not need to be complicated, and a basic one is better than none. Here is a simple seven-step approach, which is our own illustration of the method.

Define the future. List the strategic goals for the next one to three years, and identify what skills they require. Include AI-related skills and others, such as data literacy, change management and customer service.
Map current skills. Collect data on what people can do. Sources include HR records, self-assessments, manager reviews, certifications and project histories. Be clear that the aim is development, not punishment.
Compare. For each skill, estimate how many people have it, at what level, and how many you need. The difference is the gap.
Prioritise. Rank gaps by importance and urgency. Focus on those that affect key goals.
Choose actions. For each gap, decide whether to train, hire, partner or automate.
Communicate. Share the findings and plans with staff.
Review. Repeat the audit regularly, since needs change.

On-the-job training after the skills-gap audit

Workhuman’s second step is on-the-job training. This means learning by doing real work, with support.

It has big advantages. The learning is relevant, since it applies to actual tasks. It is immediate, so people can use new skills at once. It costs less than external courses, and it keeps people productive during training.

Plan it. Pick tasks that stretch people but are manageable. Provide guidance, such as a mentor or a checklist. Set clear goals and give feedback.
Make time for it. If people are too busy, training will be squeezed out. Managers need to protect time and adjust workloads.
Record what is learned. A simple log of tasks completed and skills practised helps people see progress, and helps HR update skills data.

For AI skills, on-the-job training might mean using an AI tool for a real task, such as drafting reports or analysing data, with an experienced colleague. For example, a finance team might trial an AI tool to prepare monthly summaries, then review the results together, noting where the tool helps and where it makes mistakes.

On-the-job training works best with other methods, such as courses for theory and peer learning for support. Our guide on upskilling for AI shows how technical and soft skills fit into that mix.

Peer learning after the skills-gap audit

The third step in Workhuman’s guide is peer learning. This means colleagues teaching and learning from one another.

Peer learning is powerful. People often learn best from those who do similar work, who understand the context and use language that makes sense. It spreads knowledge across the organisation, builds relationships and trust, and recognises the expertise that already exists.

Ways to do it include lunch-and-learn sessions, communities of practice, buddy systems, shared documents with tips, short demonstrations in team meetings and internal forums. In AI, early adopters can share what works and what does not.

Give it structure and recognition. Without structure, it may fade. Set regular times, a facilitator and a simple format. Recognise people who share their knowledge. Workhuman’s own business is recognition software, so it is no surprise that it stresses this, but the principle is sound: people repeat behaviour that is appreciated.
Include everyone. Make sure peer learning is open to people at different levels and in different roles. Encourage cross-team sharing.
Watch quality. Peers may share outdated or incorrect information. Have experts review key materials, and encourage people to check facts.

Peer learning is a low-cost, high-impact way to build skills, especially in fast-changing areas.

Pair upskilling with reskilling

Workhuman advises pairing upskilling with reskilling. The two are related but different. Upskilling deepens skills in a current role. Reskilling moves someone into a different one.

Why pair them? Because needs change at different speeds. Some roles will change in content, such as an analyst who now uses AI tools. That calls for upskilling. Other roles may shrink, while new ones grow. People in shrinking roles may need reskilling to move into growing ones.

Doing both gives flexibility. You can strengthen people in their current jobs and prepare others for new paths. It also shows employees that the organisation cares about their futures, which supports retention and engagement.

Plan the paths. Use the skills-gap audit as your guide. For each at-risk role, identify neighbouring roles that use similar skills. For example, a customer support agent could move into data annotation, quality review or an AI operations role, with some training.
Be honest. If roles are likely to shrink, tell people early and offer support.
Link to hiring. Before hiring externally for a new role, look at internal candidates who could be reskilled.
Measure results. Track how many people move, how well they do and how long they stay.

Pairing the two builds a more resilient workforce.

💡 Did You Know? Most Reskilling Pilots Succeed Small, Then Stall at Scale

Workhuman flags a hard truth: many reskilling pilots work well with small groups but fail when extended to thousands of people. The fix is not a bigger programme. It is shared skill categories, trained managers and a patient, step-by-step rollout. Another bonus of a skills-gap audit is that it often uncovers hidden strengths, such as languages, coding or project management that staff learned outside their job titles.

The challenge of building reskilling frameworks that scale

Workhuman lists a challenge: building reskilling frameworks that scale. Many pilots succeed with small groups but fail when extended to thousands.

Why is scaling hard? Different roles need different skills. Local managers vary in support. Budgets are limited. Data on skills is incomplete. Tools may not connect. Employees have different needs, from beginners to experts.

Standard frameworks, local flexibility. Define common skill categories and levels, so that everyone uses the same language. Allow teams to choose how to meet them.
Use technology wisely. Learning platforms and skills databases can help, but they need good data and design. Do not let tools drive strategy.
Build manager capability. Managers are the link between plans and practice. Train them to support learning, discuss careers and allocate time.
Start small, then expand. Test with one or two functions, learn from the experience and adjust before rolling out widely.
Measure and report. Track participation, skill gains, moves and business effects. Share results to keep support.
Be patient. Scaling takes years, not months.

The challenge is real, and acknowledging it is better than pretending that a single programme will solve everything.

AI-driven skills forecasting as an emerging tool

The guide mentions AI-driven skills forecasting as an emerging tool. This means using AI to predict which skills will be needed and which will fade.

Such tools analyse data from job postings, internal records, market reports and more. They can spot patterns, such as rising demand for certain skills, and suggest actions. For a large organisation, they can save time compared with manual analysis. Instead of waiting for gaps to appear, you can anticipate them, and match employees to development paths based on their skills and ambitions.

There are risks:

Forecasts may be wrong, since skills demand is hard to predict.
Models can reflect biases in the data, for example by favouring groups that have been advantaged in the past.
They may rely on data that is incomplete or out of date.
Over-reliance can lead to rigid thinking.

Treat forecasts as inputs, not as answers. Combine them with human judgement, and check them against real hiring and performance data. Be transparent with employees about how data is used, and follow privacy and anti-discrimination rules, including those on AI in employment decisions in some US states.

Because the tool is emerging, the guide gives few details, and it should not replace a human skills-gap audit. Test vendor claims, and ask for evidence.

What workers can do to take part in reskilling

Reskilling is not only the employer’s job. Workers can take an active part.

Know your skills. List what you can do, including skills from outside work. Ask for feedback.
Know the direction. Learn about how AI is changing your field. Read, ask and observe.
Speak up. Tell your manager what skills you want to build and what roles interest you. Many employers cannot guess.
Take opportunities. Volunteer for projects that use new tools. Join peer learning groups.
Learn outside work. Free and low-cost courses exist. Choose reputable ones, and apply what you learn.
Document your growth. Keep a record of new skills, projects and results, for your CV and for performance talks.
Build relationships. Mentors and sponsors can guide you and open doors.
Look after yourself. Change can be stressful. Set realistic goals, and seek support when needed.

If your employer offers no help, consider whether your role has a future, and think about your options. In a slow market, plan carefully, but do not wait until you are forced to move. If you are weighing a structured route into a new field, our explainer on internships and learnerships in South Africa is a good place to start.

How to read Workhuman’s skills-gap audit guide with care

Workhuman sells recognition and engagement software, so its guide serves marketing as well as education. The steps it proposes are widely accepted in workforce planning, but compare them with other sources.

It gives only the main steps and no data on results, such as how many organisations have succeeded with this approach. It is a how-to rather than a research report, so it offers advice, not evidence.
The publication date, around 14 October 2026, is approximate.
Terms such as reskilling and upskilling are used loosely in the market. Check definitions.
The guide is aimed at employers in general, mostly in the US. South African employers should adapt it to local conditions, including skills development legislation, learnerships and levies. For instance, employers may be able to link training to sector education and training authority programmes. Check current rules.

Finally, keep in mind that no single programme guarantees success. Reskilling depends on leadership, resources, culture and the willingness of people to learn. Use the guide as a starting framework, and adapt it.

Quick Revision: Skills-Gap Audit Flashcards

What does Workhuman say to do first when reskilling for AI?

Run a skills-gap audit

Which two learning methods follow the audit?

On-the-job training and peer learning

What should you pair with upskilling?

Reskilling

What emerging tool does the guide mention?

AI-driven skills forecasting

Frequently Asked Questions About Skills-Gap Audits and Reskilling

What is a skills-gap audit?

A skills-gap audit compares the skills your organisation will need in the near future with the skills it has now. The difference between the two is the gap, and it tells you where to train, hire, partner or automate.

Why should a skills-gap audit come before training?

Without an audit, training can be scattershot: people take courses that do not match business needs while real gaps stay open. Starting with an audit saves money and time, builds a shared picture and supports fairer decisions.

What is the difference between upskilling and reskilling?

Upskilling deepens skills in someone’s current role, such as an analyst learning to use AI tools. Reskilling prepares someone for a different role, such as moving from a shrinking job into a growing one. Workhuman advises pairing the two.

How do on-the-job training and peer learning help?

On-the-job training is relevant, immediate and low cost because people learn by doing real work with support. Peer learning spreads knowledge across teams through buddy systems, communities of practice and short demonstrations. Both work best when they are planned and recognised.

Why is it hard to scale a reskilling programme?

Roles need different skills, local managers vary in support, budgets are limited, skills data is incomplete and employees have different needs. Standard frameworks with local flexibility, trained managers and a start-small approach help.

Can AI predict which skills will be needed?

AI-driven skills forecasting is an emerging tool that analyses job postings, internal records and market reports to spot rising and fading skills. Forecasts can be wrong or biased, so treat them as inputs and combine them with human judgement and a proper skills-gap audit.

Final thoughts

A skills-gap audit is the right first step in reskilling for AI. Workhuman’s guide then advises on-the-job training, peer learning and pairing upskilling with reskilling, while noting the challenge of building reskilling frameworks that scale and the emerging role of AI-driven skills forecasting.

The practical path is to run a skills-gap audit, prioritise, train through real work, share knowledge among peers, plan both upskilling and reskilling, and review regularly. Workers can take part by naming skills they want, taking opportunities and documenting growth. Employers should be honest, patient and measured in their approach.

Sources and further reading

Workhuman, Reskilling in the Age of AI: How to Future-Proof Your Workforce (published around 14 October 2026)

Summary

Skills-Gap Audit: Workhuman's Guide to Reskilling for AI