AI skills planning: Practical guidance for people professionals

Clearly, the impact of AI on skills will be huge. The skills required for some jobs will change as employees start using AI to do things they currently do themselves. For example, graphic designers will still need creative design skills, but as they use AI to do more of the actual creation, the focus of their people skills will shift towards interpreting consumer reactions and client needs, as the nature of their job evolves. Other jobs will need new skills, for example doctors will no longer need to analyse scans to spot cancer, but they will need new skills to verify the work of the AI that now does that for them. One aspiration is for AI to take away some of the repetitive and mundane tasks, freeing up people to focus on areas that add greater value to the business or its customers.

The goal of AI skills planning is to align people changes with AI adoption in a co-ordinated way. Done well, it enables a smooth transition from the early stages of experimenting with AI, to the advanced level of AI being fully integrated into human work. If done poorly, HR faces a constant ‘catch-up’ – reacting to new uses of technology that suddenly need employees to quickly acquire new skills in order to change the way they do their current jobs or perhaps carry out completely new jobs.

The key is recognising that AI works its way into organisations in stages – referred to in this guide as the three levels of ‘AI adoption maturity’ – and being prepared for each stage. This isn’t necessarily something business or IT leaders plan – it’s just the way transformational technologies become part of business. The difference with AI is the speed of technology adoption. HR had years to prepare for and manage the job and skills impact of computing, the internet and mobile communication. With AI, the change happens in months and quarters, and HR must adapt to this new, unfamiliar pace by responding more dynamically than traditional annual planning cycles typically allow.

As you reach each level of AI adoption, you should ideally have a picture of the human skills your organisation will need for that degree of AI use. The AI skills plan is your map for moving up through the three levels of AI adoption as a managed process, which allows your business to get the most out of its AI investment with the least human disruption.

To navigate between levels, you’ll need to know what skills you currently have and which ones you’ll need for the next level of AI adoption. For the changes and new skills, you’ll have to decide whether the answers will lie in training existing staff, hiring new people with different skills, or moving the skills outside the employee base, eg outsourcing or contractors.

Alongside these skills changes, you’ll need to systematically manage the job changes and the associated impact. These will range from job descriptions and performance management to organisational restructures and career paths. And, of course, quite possibly some redundancies.

In other words, AI skills planning is one piece of an AI adoption change management journey. So good organisational change management practices will be needed, like effective employee communication, visible leadership and effective metrics. Part of the AI change programme also includes ensuring employees have a basic level of digital literacy, as a prerequisite to adopting AI.

Below you’ll find a five-step structure that covers all these, helping you identify, plan and manage the human skills changes you’ll need for your organisation’s next level of AI adoption. This structure is designed to allow skills planning to slot into any bigger AI change management programme you may be following. But even if that’s not how your organisation is treating AI adoption, you can still use the five steps as a standalone piece of HR work.

AI adoption journeys are different for each organisation, reflecting unique strategic priorities, operating models, industry practices and risk context. However, there are common patterns, challenges and pitfalls that occur. The information below is designed to help people professionals anticipate and plan for the workforce capability implications of AI adoption, supporting skill development in a deliberate, sustainable and strategically aligned way with your organisational goals.

The starting point is deciding your AI adoption level, so you know the leg of the journey you’re on: early stage to intermediate, intermediate to advanced, or advanced to beyond.

Familiarise yourself with the generic AI skills planning process below, and review how it will fit against how you currently do skills planning in general. 

• If you already manage skills using an approach like strategic workforce planning or competency development frameworks, you can use it as the basis for AI skills planning. You’ll need to map the five steps described below against your own process, and integrate them appropriately for your circumstances.
• If you don’t do skills planning formally, but want to, you could use this AI skills planning process as a starting point for a more systematic approach. The CIPD can provide guidance on selecting an appropriate approach for your circumstances.
• If you’re not ready to do skills planning formally yet – like many organisations, mostly SMEs – think seriously about doing so just for AI skills, and possibly digital skills as well.

Once you’re comfortable with the high-level process, you can go to the section describing the detailed process for your level of AI adoption. There are three options: one for early-stage AI adopters, one for intermediate AI adopters, and one for advanced AI adopters.

Use the self-assessment quiz to find out your level of AI adoption maturity.

Go to the appropriate AI skills planning section for your level:

• AI skills planning for early-stage AI adopters.
• AI skills planning for intermediate AI adopters.
• AI skills planning for advanced AI adopters.

Details of each of the general steps are described below.

1. Align (The “why” and “what”)

Strategic alignment is about ensuring AI skills work serves a business purpose rather than just following technology trends. Its aim is to build skills plans around what the business is trying to achieve through the technology, as well as how to operate the technology. This requires input from technology colleagues on specific AI technology that will be used, for example ChatGPT or Copilot.

HR must partner with business and tech leaders to define the ‘why’ of AI. For example, is the potential aim of AI to enhance products (eg, get better customer insights about their needs and wants) or to drive efficiency (eg, automating complex data entry)?

Establishing timescales is crucial, so get a sense of anticipated management time horizons for AI. An aspiration to fully automate a function in six months needs a much more aggressive reskilling plan or transition strategy than a three-year pilot programme.

2. Audit (The 'as-is')

Auditing is about uncovering the current reality of AI use. This often reveals ‘hidden’ AI users, such as employees using tools for specialist tasks through their own initiative, without formal oversight – for example to speed up data analysis or report writing.

Beyond technical ability, the audit should gauge employee attitudes. Identifying where there is enthusiasm for AI can help you find ‘AI champions’, while uncovering fear allows you to tailor your communication and change management to address resistance early.

3. Analyse (The 'gaps')

The goal here is to identify the distance between your current audit results and your future business goals. This involves distinguishing between the need for high-volume general AI literacy and low-volume, high-impact technical skills.

Be pragmatic: prioritise gaps that pose the highest risk (like lack of AI ethics knowledge) or offer the quickest wins (like prompt engineering for high-volume writing tasks). Document your assumptions about when these skills will be needed to ensure training isn't delivered too early or too late.

Be sure to also consider the impact of AI on non-AI skills. For example, the adoption of AI to generate written documents will reduce the need for skills associated with document preparation, such as effective business writing and preparing executive briefings.

4. Act (The 'how')

Once gaps are identified, HR must decide between four strategies: buy (hiring new talent), build (training existing staff), borrow (using consultants), or bot (automating the task entirely).

While ‘borrowing’ talent is fast, it doesn't build long-term internal capability. ‘Botting’ is the most complex from a people perspective, requiring deep collaboration with IT and finance to manage the fallout of automation, such as potential redundancies or the need for significant role redesign.

In addition to the operational aspects of filling the skills gaps, it is equally important to win hearts and minds so that the gaps stay filled. Unlike other forms of skills planning, AI-related HR work has to deal with unprecedented degrees of employee scepticism and concern, so the softer aspects of organisational change such as confidence building are an integral part of AI skills planning.

5. Adjust (The 'evolution')

AI is moving so fast that an annual skills plan is likely to be obsolete before it is finished. Shorten review cycles, ideally quarterly, to regularly reconfirm that the AI tools and new skills you planned for are still appropriate and your risk assumptions are still valid.

You should also plan for ‘crisis’ skill needs – having a ‘break glass’ plan to access expert AI help if a system fails or leads to errors that cause financial or reputational damage.

AI skills planning will be more successful if you're armed with useful metrics, especially ones that can be linked to financial benefits and costs. These will have a variety of uses beyond skills planning, from monitoring change management progress to securing budget for HR initiatives. 

However, the metrics that will be useful for AI skills planning are different to other business metrics. Some are AI variations of familiar HR measures – like training completion rates. But others are completely new – like hallucination tracking. Below are a few typical AI skills planning metrics, to illustrate what you’ll be looking for in your organisational circumstances.

Read the statements below and choose one answer per question that most closely describes your own organisation today.

If you fall between two answers, or aren’t sure, err on the side of the less mature of the two descriptions that might apply. 

Obviously, this is a very simplified assessment of AI adoption maturity, boiled into three questions solely to help use this guide. Assessing AI adoption maturity more fully requires a more sophisticated approach covering aspects like what data you have and how you collect it, aspects of organisational culture like attitudes to experimentation and risk, and how you approach ethical questions.

If you’d like to explore this further, the CIPD can help you find guidance on AI adoption maturity frameworks and methods and connect you with organisations with appropriate expertise like The Alan Turing Institute.

Scoring

Mostly A: Early-stage: Use the early-stage AI skills planning.

Mostly B: Intermediate: Use the intermediate AI skills planning.

Mostly C: Advanced: Use the advanced AI skills planning.

• Identify AI champions: Look for ‘hidden’ users already experimenting with tools like ChatGPT or Gemini.
• Establish a ‘sandbox’: Create a safe, approved space for employees to ‘play’ with AI on company time without risk to data privacy.
• Run an ‘interest’ audit: Instead of testing for mastery, survey staff to see who wants to learn about AI.
• Set ‘safety-first’ rules: Publish light-touch guidelines on allowed tools and how to protect company data.
• Curate free resources: Share high-quality, low-cost webinars or YouTube videos for non-technical users. 

The aim of this step is to provide an initial base for skills planning, on the understanding that this may well change, in which case the skills needs will change. The emphasis of strategic alignment when AI is still new and unknown is turning curiosity and uncertainty into provisional decisions and working assumptions. This includes raising awareness that people may already be using AI confidently without realising it, for example when searching the internet in modern browsers, or getting help while shopping online.

Business and technical leaders are unlikely to have clear or evidence-based answers to questions about the ‘why’ and ‘what’ of AI for the organisation. But they may feel pressure to come up with answers, especially in response to competitor, customer or other business stakeholder pressure.

For AI skills planning purposes, it may be enough to identify some areas the organisation feels appropriate (safe) to experiment with AI. Make sure this includes defining explicit boundaries and ‘no-go’ areas such as anything involving personal data or product safety. Even better would be reviewing all departments, or functions, and coming up with initial, tentative working assumptions about the impact of AI, to validate when appropriate. 

For beginners, the audit should be low-key, relatively informal and focus on interest rather than mastery to avoid raising concerns. Be aware of and cautious about the risk of inadvertently sending unintended messages with the act of auditing, eg signalling a strategic decision to start using AI. 

The emphasis should be on interest rather than ability – ‘who wants to try out AI?’ rather than ‘who knows about AI?’ or ‘how well can you use AI?’ But also look for enthusiasm and experience, possibly to identify ‘AI champions’ who are willing to be informal leaders in the organisation’s experiments with AI.

Catalogue the different tools people have experience of, eg Co-pilot, ChatGPT etc, and the different tasks they’ve played with, eg image editing, information seeking.

Consider uncovering employee concerns about AI, where possible specifically about skills, but other concerns may be unavoidable. This is an example of where the overlap between change management and skills planning is particularly strong, so good change practices should be used. For example, acknowledging nervousness and scepticism rather than trying to avoid it.

For this and all following sections, ‘employee’ refers to all levels of the organisation, including managers and senior leaders. Management perspectives on skills will provide crucial inputs, including the impact of AI on management and leadership skills.

The focus of gap analysis early in the AI adoption journey is skills that minimise risk and inadvertent harm, rather than maximise business benefits and efficiency. Don’t ‘run before you can walk’, so don’t prioritise skills for AI in work until you’re confident affected employees can use AI safely. 

This means prioritising digital literacy, knowledge of basic AI concepts (eg, what is Gen AI?), data security, and what tools are employees allowed/not allowed to use at work.

Be cautious about creating unintentional, implicit digital divides by leaving those not involved in early AI skills activity feeling left out or left behind.

Use good practice in employee communication – such as honesty and transparency – to reassure all employees about AI, especially those expressing the greatest concerns.

Create safe, approved opportunities for employees to ‘play’ with AI on company time.

Publish light-touch rules and approval procedures for employees who want to use these ‘sandboxes’. They should provide brief, mandatory feedback for skills planning.

Organise employee access to free or low-cost AI education resources, including YouTube videos, webinars and even basic certifications in AI for non-technical users. This could include company campaigns that help shine a light on the subject and dispel myths. 

Consider including skills activity that focuses on reflection and discussion, rather than teaching and practice. Early-stage adopters have an opportunity to manage the AI change journey thoughtfully from scratch, unlike more advanced organisations.

Capture quantitative and qualitative information that will help skills-building plans and decisions, eg number of employees testing tools, attitudes to AI, key concerns.

Use employees not yet familiar with AI to uncover ‘Emperor’s New Clothes’ questions and issues, eg why they’re using AI for something a spreadsheet does just as well.

Build into your plans a way of monitoring AI adoption maturity, so that you can see the progression from early-stage to intermediate in enough time to adapt.

Consider asking employees for input on ways AI could be used to make the business better. They may see opportunities not apparent or visible to business leaders.

Track qualitative data: how are attitudes changing? What are the key concerns? Use feedback from ‘non-experts’ to challenge the hype, for example why AI is being used to do something a spreadsheet does better.

Use your data to monitor your AI adoption maturity, so you can prepare for the transition to the ‘intermediate’ stage when ready, when AI skills planning will become more formal. 

What was happening?

Store managers began using personal AI tools for shifts without oversight.

What did HR do?

HR carried out an anonymised survey to understand where and how staff were already using AI. This helped identify common examples of ‘shadow AI’ use, as well as employees who were confident experimenting with the technology. HR then worked with these employees as informal AI champions, asking them to help identify practical use cases that could be tested safely. Alongside this, HR developed a simple AI use policy, setting out what tools could be used, what data should not be entered, and when human review was required.

What changed as a result?

BrightPath moved from unmanaged experimentation to a more transparent and controlled approach. The organisation identified three AI champions, agreed priority use cases for safe piloting, and gave managers clearer guidance on how to use AI responsibly before wider rollout. 

• Standardise skills records: Move from informal notes to spreadsheets or HR system to track AI competencies. 
• Pilot task-specific prompting: Train teams in active pilot projects on how to write effective prompts and, crucially, how to evaluate the results.
• Mandate human verification: Focus on human responsibility skills to check AI outputs for 'hallucinations' or errors.
• Review supplier lists: Check if your L&D and recruitment partners can provide the AI skills you’ll need.
• Track time and quality: Start measuring time saved or quality improved in teams currently using AI.

Alignment at this stage of AI adoption shifts from individual curiosity to team-level feasibility, or, if more appropriate, function or department level. The aim is to create AI projects to pilot and evaluate the feasibility and reality of AI in the organisation. 

Start to consider measurable goals/KPIs for AI and its impact on business metrics. HR should work with department leads to establish these for each AI pilot, such as time saved on routine admin or improvements in output quality.

It is also the time to look externally at how AI is affecting the sector and competitors. But review this with healthy scepticism, to avoid being influenced by ‘hype’ rather than fact.

Move beyond individual interest to assess skills by job family, team or function, including whether these are already developing or still lacking.

Identify specific job families or roles where AI is already altering daily responsibilities, showing a need for collective rather than individual AI skills or responsibilities. 

Evaluate the effectiveness of any early learning resources, such as online courses or internal webinars. Ask employees if they have experience of other, more effective skill-building resources, eg TikTok, Coursera, YouTube.

To build future scalability and flexibility, start to formalise how you document AI skills, shifting from ad-hoc or free-form notes to structured records, eg move information in Word documents to Excel, explore if your core HR systems can capture/report skills.

Priority should be given to skills that mitigate AI risk and manage potential errors, such as ‘hallucinations’ in generative AI. If this has been addressed previously, consider refresher courses to ensure the skills have not been forgotten.

Once safety is addressed, focus on task-specific prompting (and evaluating) for teams currently in pilot phases using Gen AI, or those expected to do so.

The final core skill to prioritise at this stage is the critical ability to review and ‘human-verify’ AI-generated work. It’s relatively straightforward to start using AI in business, but much harder to recover from the problems and consequences of using it poorly. One of the biggest sources of AI difficulties is humans over-relying on the technology and using its results or following its suggestions blindly. This is particularly the case when the early experiences of using AI have been good, and it becomes tempting to just carry on using it unquestioningly. The caution and judgement to avoid this is a critical AI skill to build early.

Feel free to look at any other skills gaps you’ve identified, once these priorities have been addressed. But be cautious about management pressure to focus on skills that will allow more benefit-focused uses of AI, eg efficiency improvement ahead of these. 

In this stage, evaluate the buy vs build vs borrow vs bot options for developing skills, but focus heavily on build, engaging with L&D colleagues. Building AI skills and experience across as much of the workforce as possible will create long-term resilience and reduce dependency on acquiring external skills.

At this level of adoption, it is important that HR has explored options for professional AI training providers, because of dependency on procurement, eg changing supplier evaluation criteria to reflect AI skills needs, and perhaps getting new suppliers onto approved supplier lists.

This is less critical if you have relevant in-house or freelance training resources already in place, but it is still better to have third party suppliers available in place if they are needed in future. A collateral benefit is the knowledge you’ll gain by simply discussing your AI skills needs with training companies, as they are likely to have useful insights based on what other customers are doing and asking for to deal with AI skills. 

In parallel, review recruitment needs for specialist skills that cannot be developed quickly in-house, or which are not part of long-term reskilling plans.

This is also the stage to update HR policies, such as data security and acceptable use, to reflect formal AI integration. Depending on organisation culture and priorities, this may be the right time to start discussions about the environmental impact of AI, linking AI adoption to organisational sustainability policies. This can be linked to AI skills planning through training individual users on minimising their personal AI environmental footprint, for example through how they construct their ChatGPT or Copilot queries.

Finally, start considering how you would deal with any outsourcing or automation triggered by AI, eg potential redundancies, employee communication, setting KPIs etc.

This is the stage to start tracking the business impact of pilot projects alongside technical performance, for example time saved, quality improvements etc.

Also ensure AI errors in pilot outputs are tracked, eg hallucinations, quality problems etc. As part of looking at AI errors, this is also the time to start considering some of the ethical challenges of AI, such as the risk of bias and discrimination in people decisions. This is a precursor to more systematic work required later on AI fairness, for example to ensure existing diversity and inclusion issues are not replicated in algorithms.

It is vital to monitor employee sentiment during this phase, both positive and negative, to ensure the workforce feels supported during this transition. Use these insights to refine training programs before they are rolled out more widely across the organisation. As part of this monitoring, ensure that employees who may be disadvantaged feel they have as much access to AI as others. For example, pilots should not be restricted to those with pre-existing technical confidence, or withheld from those with an openly negative attitude to AI. 

What was happening?

A department-wide pilot using AI for route optimisation showed high ‘hallucination’ rates.

What did HR do?

During the audit, HR created a simple skills-tracking template to understand where staff needed support. They prioritised training that helped staff critically assess AI-generated outputs, spot errors and challenge inaccurate recommendations before they reached customers. They also worked with operational leads to develop task-specific prompts and examples of good practice, helping teams use the tool more consistently and safely.

What changed as a result?

Error rates dropped by 40%, and the team successfully transitioned to ‘task-specific prompting’ to refine results.

• Embed AI in job descriptions: Make ‘AI skills’ a mandatory category in all role descriptions, even if the current requirement is ‘none’.
• Draft a workforce restructuring plan: Lead senior management discussions on which roles will disappear, merge, or be created due to AI.
• Launch an AI Academy: Establish a formal in-house learning hub or partner with an external specialist for structured certification.
• Redesign career paths: Map out new progression routes for junior staff, particularly where ‘entry-level’ tasks have been automated.
• Integrate AI ROI: Link AI skills data with financial performance metrics to track the true return on investment. 

The skills impact of AI on advanced AI adopters is strategic and structural, and needs to be managed at senior-most levels. Good AI skills planning – and indeed good AI adoption – should be driven by evidence and clear business metrics. 

HR therefore must actively participate in board-level discussions and decision-making about strategic plans for the use of evolving AI, not just from a people perspective.

The ‘why?’ and ‘what?’ questions about AI now need to include ‘so what?’ discussions that address the hard choices and consequences with people impact.

In particular, HR should be leading – or at least strongly influencing – senior management discussions about workforce restructuring. This must specifically include how to deal with roles that will need to disappear, merge or be created. Unionised organisations have, of course, additional complexities and obligations on this.

Strategic AI adoption should be treated as an organisational change programme, including good practices for stakeholder engagement and communication.

Tracking AI skills now needs to be formal, scalable and integrated into regular HR processes such as performance management and personal development plans.

AI skills should be included as a mandatory category of all role/job descriptions, even if blank, ie even for roles that do not (yet) require AI skills.

Consider creating an AI skills impact document for each function, department or team, that covers the future as well as present. Ideally, provide support for line managers to prepare these themselves, as part of building their AI management skills.

Pay particular attention to AI governance and ethics skills at senior levels, as these will become increasingly important business responsibilities for managers and leaders. 

The focus of skills planning now shifts from ‘avoiding harm’ to ‘increasing efficiency’ or ‘improving performance’. As with the earlier stages, this is not about AI skills in isolation, and is intrinsically linked with other aspects of change, including digital confidence and capability.

Use business metrics and risks to identify and prioritise the skills gaps that will allow AI to deliver benefits. The earlier emphasis on skills to avoid AI harm should no longer drive skills planning, but must not be overlooked completely. These still remain critical. 

Identify leadership skills needed for widespread AI use, particularly new ones such as managing hybrid AI-human teams or resolving ethical issues not considered before. Part of this assessment is the attitudes of line managers, to assess how much hearts and minds work should be planned for to win them over.

Pay attention to the need for cross-functional collaboration, and any skills or role impact, eg legal inputs on new AI regulation, IT co-operation to analyse AI errors.

Redesign career paths and progression options around a new view of the organisation. Create a picture of the organisation – such as a Target Operating Model or similar – which has AI embedded strategically into processes and is likely to increase both the use of AI and the power of the AI technology used over time.

This will allow you to do some form of scenario planning and ‘what-if’ analysis from a people viewpoint. It will uncover difficult long-term issues you should prepare for now. For example, deciding how you will deal with the removal of entry-level and junior roles as AI automates that work – those employees form the pipeline for mid- and senior-level roles. 

Consider an AI academy for all AI skills work, either in-house or partnering with 3rd parties. Alternatively, provide employees with structured AI certification opportunities.

Find specialist partners who can provide senior leadership education on evolving AI governance, and help the organisation decide and design its own responses.

Consider using AI auditors to independently review your organisation’s response to AI, especially risks and compliance. Avoid superficial ‘tick-box’ AI certification firms.

Practitioner tip: Skills vs. work design

While this guide focuses on skills, remember that as you ‘bot’ or automate tasks, the nature of the work itself changes. You must actively redesign roles to ensure that remaining human tasks form a meaningful, coherent job. Simply training someone for a new skill is ineffective if the surrounding role structure is broken.

AI skills monitoring should be integrated into normal HR skills planning, using regular HR processes and systems, eg workforce analytics to capture relevant information.

Business metrics should be in place to track benefits of using AI, ideally as part of AI ROI tracking. AI skills should link to these in some way, depending on circumstances.

The increasing use of ‘AI + human’ teams will mean the cost of AI ROI becomes much more complex to calculate. HR should ensure the people (and skills) component is not separated out inappropriately, to avoid errors in evaluating the financial impact of AI. 

As job cuts and displacement become more likely, tracking employee sentiment will become increasingly crucial. This should at minimum be part of AI skills monitoring, but ideally should be moved to a broader AI change management programme.

What was happening?

Strategic plans to embed AI in diagnostics threatened to remove entry-level junior roles.

What did HR do?

HR led board-level discussions on workforce restructuring. They launched an internal AI Academy and redesigned career paths so junior staff could learn ‘AI auditing’ skills instead of manual data entry. 

What changed as a result?

The firm achieved a measurable AI ROI while maintaining high employee sentiment by being transparent about job evolution. 

Be clear and honest with employees about why AI skills are being built. 

Be ready with management positions on difficult questions, eg automation, job losses, ethical issues and environmental impact.

Have a policy – ideally published – on support for tasks that will be replaced or require significantly changed skills.

Use HR communication best practice to guide how you do this, eg how to deliver difficult messages, what to do when answers don’t exist yet etc.

Ensure access to training is inclusive, ie not restricted to certain grades or roles, and without the 'volunteer bias,' of only tech-confident employees participating. Be sensitive to management bias, such as assuming older workers are less interested in AI upskilling. 

If budget prevents that, then ensure the allocation of training is equitable. Be sensitive to the possibility of senior management bias on this, eg age discrimination.

Be aware of AI skills gaps creating an in-house ‘digital divide’, where some groups are left behind with AI skills. Monitor for this, using the AI skills audit as a baseline.

This is about ensuring employees realise they are accountable for AI outputs, actions and decisions. It won’t be acceptable to say ‘the AI did it’. 

Incorporate the implications of this into high priority skills assessment and training, ie critical thinking in the context of evaluating what AI does and says.

Experimentation and trial-and-error should be encouraged, but only with understanding and appreciation of risks and possible negative consequences. This will only be possible if managers create the right psychological environment, where reassurance about the consequences of making mistakes will be perceived as authentic.

Ensure skills and training for those using AI – particularly those experimenting or working without supervision – covers data privacy and other active risk mitigations.

Consider creating ‘safe’ areas in which employees can practise and learn using AI for work tasks without risk, eg actively monitored/supervised, test environments etc. 

Use internal data such as results of skills audits and pilot projects as the basis for skills planning. Avoid using external data – such as industry statistics – unless you can be sure of its validity and relevance. 

Do not use external hype in your AI skills plans. Be ready to push back if others in the organisation do so – especially senior leaders deciding the why and how of AI.

Regularly review the skills frameworks and technology assumptions you’re using in your AI skills plan – ideally at least every six months.

If people professionals are to successfully guide the rest of the organisation on how to adopt AI – specifically around skills – they will need enough credibility to be listened to.

While there should be no question about the people profession’s authority in skills in general, there may questions around AI knowledge. Whether accurate or not, employee perceptions of AI may well be that it’s an IT problem, even though the reality is far more complex, as we’ve seen so far in this guide. 

So when you stand up in front of colleagues to talk about how to deal with AI, you should be ready to convincingly answer the question ‘But what do you know about AI?’ Unfortunately, our research suggests that all too often, the answer is ‘not that much’, in large part because HR frequently hasn’t actually used AI for themselves much in their own work. This can often be accompanied by a diffidence that comes from discomfort with technology, which won’t help your cause. You can probably help yourself with some non-technical training about the basics of AI, but there’s no substitute for experience.

People professionals should seriously consider taking responsibility for avoiding credibility challenges, by forcing themselves to experience for themselves the natural human discomfort of discovering that a computer appears to be able to do some parts of your job for you already, often faster or even better than you could. Again, the reality isn’t so simple, but you need to go through the uncomfortable experience of seeing that in practice for yourself, then coming through to the other side and realising that isn’t necessarily as bad as it initially feels, and is actually an opportunity. For example, freeing up time from ‘gruntwork’, or treating AI as a tool to help you get more done in a day rather than replace you.

You may already know these things intellectually, but until you’ve experienced what they mean personally, you risk being emotionally disconnected from your employees when you lead them through the skills impact of AI. The most authentic answers to questions about what it means to them will include your descriptions of what it meant to you.

It’s now well established that AI will fundamentally reshape how talent moves through organisations, by disrupting our existing models of career paths. The existing principle of senior people ‘working their way up’ and learning skills along the way doesn’t apply if the starting jobs at the bottom of the organisation don’t exist any more – at least in the same form. And if those jobs are completely replaced (botted), employees will no longer have the skills and experience on which to build what they need higher up the organisation.

Another consequence of more AI in the workplace is on management skills. Specifically, the current skills needed to manage human teams will need to be revised and perhaps even re-imagined for teams of humans equipped with supercharged AI tools, or hybrid teams of humans and AI agents working together.

The immediate challenge for HR will be getting ahead of this by redesigning career paths now, for a future that may not happen for a while. From an HR perspective, that future will need to give junior employees alternative ways to learn what they will need as senior employees, in the absence of on-the-job learning and experience. 

When we touched on the discomfort of experiencing the impact of AI on your own job earlier, we didn’t talk about the potential impact on a person’s emotional wellbeing and mental health. The modern workplace should include mechanisms for providing mental health support, but AI is likely to test how well they work, and if they’re sufficient for the kind of problems some employees may face. While the concerns are familiar – job security, fear of change, concern about being too old to learn new skills – the triggers are different. We can learn from previous generations of comparable change, such as outsourcing and process automation, but the lessons need to be brought up to date. 

The key is a two-fold approach. Firstly, make sure your support mechanisms are in place, and are adequate for the spikes of demand that may follow announcements and decisions about AI. But also look for ways to minimise or avoid AI anxiety in the first place. For example, we’ve mentioned in the guide prioritising safety-first training and experimentation for that reason among others. Employee wellbeing is also something to consider including in the measures and monitoring your set up as part of AI skills planning.

The final consideration for HR during skills planning is a major one, but we’re only going to touch on it briefly. That’s because there’s so much to cover that this guide isn’t the right place to discuss it meaningfully. The topic is what the role of HR is and should be in wider AI strategy and governance.

In essence, HR needs to be an active, visible and influential voice in organisational decisions about what to do with AI strategically, and how to control and manage AI safely and securely. People professionals are needed there because they are the only ones in most organisations who both deeply understand the human perspective in business, and care about it as part of their professional responsibilities.

How HR gets itself into that position, and makes itself heard, is the subject of another conversation, but one that the CIPD is deeply interested in having with you.

CIPD resources

CIPD AI hub

This is the 'go-to' web page for the CIPD’s AI resources. From here you’ll find our research, webinars and guides about AI, curated for people professionals.

CIPD 'AI in the workplace' discussion forum

A peer-to-peer space where HR practitioners share real-world experiences and knowledge about AI – particularly relevant for early-stage and intermediate AI adopters.

CIPD Profession Map

This now includes an AI layer of Core Knowledge benchmarks. These are the specific topics people professionals need to know about AI, for different levels of HR work.

Other resources

The Alan Turing Institute – Bridge AI Support

Bridge AI is a UK government-funded programme to drive the adoption of AI. This includes helping UK businesses with AI upskilling, which is delivered by The Alan Turing Institute.

UK Government (DSIT) – AI Skills for the UK Workforce

The Department for Science, Innovation and Technology published this report on AI skills UK businesses will need to be AI-ready, including digital literacy in general.

FutureDotNow

This coalition of organisations is working to close the 'essential digital skills' gap present across the UK workforce, and provide a range of resources and events to aid this.  

AI Skills for the UK Workforce (UK Government – DSIT)

What is AI Skills for the UK Workforce?

A practical roadmap to identify and develop AI competencies, including how to manage and control it (governance). It maps skills across three levels: entry, mid, and managerial.

When to Use AI Skills for the UK Workforce

Particularly effective for 'early-stage' or 'Intermediate' adopters, providing a structured way to define what 'AI literacy' means for different roles, including technical and non-AI.

How to Use AI Skills for the UK Workforce

Use the Employer AI Adoption Checklist to build a detailed understanding of your organisation’s AI adoption level and maturity, and the use-case personas to test and validate it.

The Profession Map (CIPD)

What is the CIPD Profession Map?

The benchmark for people professionals, with values, knowledge, and behaviours needed as HR careers develops. It includes AI as one of the core areas of required knowledge.

When to use the CIPD Profession Map

Use this when looking at AI skills of your own HR teams, particularly during Align and Audit. You can use the AI benchmarks as the basis for your own AI skills plans in HR.

How to use the CIPD Profession Map

Treat AI literacy as either a 'specialist knowledge' area or an 'overlay' across other core HR benchmarks like 'evidence-based practice'. Either way, integrate it into your plans.

AI Skills for Business Competency Framework (Alan Turing Institute)

What is the AI Skills for Business Competency Framework?

A comprehensive tool to help define competencies required to work effectively with data and AI. Focus is not just on technical functions and roles.

When to Use the AI Skills for Business Competency Framework

Use this when your organisation moves into the 'intermediate' or 'advanced' stages and needs a more rigorous, systematic approach to defining technical roles.

How to Use the AI Skills for Business Competency Framework

Map your existing job families against the framework’s competency levels. This will help you identify gaps in technical skills and roles, and focus on those with greatest impact.

Essential Digital Skills - FutureDotNow

What is Essential Digital Skills?

Defines the digital skills everyone needs for life and work, now updated to integrate AI skills e.g. AI literacy, prompt engineering into its existing categories and structure.

When to Use Essential Digital Skills

Use for the 'Foundational' layer of AI skills planning, especially if you have a 'digital divide' across employees. Helpful for early-stage adopters and AI-diffident employees.

How to Use Essential Digital Skills

Benchmark AI skills using the '20 Tasks for Work'. Identify how to integrate these into existing L&D mechanisms, eg induction programmes, performance review processes.