AI skills in workforce planning: a worked case from one resignation to a one-page plan

Part of our topic guide on AI Skills for Business.

By James Cotton, Founder, iO-Sphere

The resignation, and the advert ready to go out

Picture, as an illustration, a distribution business in England of about 300 people. About a third work in the office, in customer service, finance, HR and a small operations team; the rest are drivers, warehouse staff and vehicle technicians. The firm has no written workforce plan. Its HR administrator has just resigned, and the old advert is ready to go out again.

In the CIPD's Resourcing and talent planning report 2026, 46% of HR professionals agree with the statement "My organisation tends to take an ad hoc approach to recruitment, replacing employees when they leave, without reviewing current or future requirements". YouGov surveyed 1,014 UK HR professionals responsible for resourcing and talent planning online from April to May 2026, weighted to represent UK employers by size and sector.

In the same survey, nearly half agree that their organisation has a workforce planning strategy based on a robust understanding of current and future workforce needs (49%), and that its recruitment is connected to a wider talent management strategy (47%). The report recommends a shift "from reactive hiring to skills-based workforce planning".

CIPD's workforce planning factsheet, dated 29 July 2025, defines the work as balancing the supply of skills against the numbers needed, and says it "doesn't need to be complicated". For a firm bringing in AI, a vacancy is the moment AI skills enter the plan or drop out of it. So this firm holds the advert and starts with the one role in front of it.

Reading the firm's roles against the exposure research

Before the role is rewritten, the HR director reads the firm's jobs against the official research on how exposed work is to AI. CIPD's factsheet puts this at the start: its first planning stage asks about productivity plans, including "plans to introduce or update technology".

In November 2023 the Department for Education published The impact of AI on UK jobs and training, one of the first attempts to quantify AI's effect on UK jobs. It scored occupations by how far the abilities they need could be aided by 10 common AI applications, a measure of exposure that did not distinguish jobs AI would aid from jobs it would replace.

The report placed human resources administrative occupations on both of its most-exposed lists, 15th for all AI applications and eighth for large language modelling. It also put them among 16 occupations labelled "high automation", drawn from International Labour Organization research, beside bookkeepers, payroll managers and wages clerks, call and contact centre occupations and customer service occupations.

That covers the firm's HR, finance and customer service work. Human resource managers and directors came 17th on the language-modelling list, and the report said its rankings "should be interpreted with caution". It found the least exposed were employees "with qualifications at level 3 or below in building and construction, manufacturing technologies, and transportation operations and maintenance".

Skills England updated the 2023 analysis in its annual skills report 2026, published in June 2026. It found exposure highest in professional, analytical and higher-paid occupations, where tasks align with what today's AI systems can augment or perform: "cognitive, clerical and data driven activities". It said sectors "centred on physical activity or human interaction, including construction and hospitality, remain less exposed".

The firm judges where its drivers, warehouse staff and vehicle technicians sit from what each job involves. The office roles are clearer. The HR administrator's job is on both of the 2023 report's most-exposed lists and in its high-automation annex, so the HR director re-scopes the vacancy before any advert goes out.

Where each skill will come from

With the vacancy re-scoped, the firm has to choose how it gets each skill the role now needs. In a CIPD opinion article in April 2025, Peter Cheese, then the CIPD's chief executive, wrote that businesses which understand their current and emerging skills "will be best placed to 'buy, borrow or build' the skills they need". Employers, he added, "must now also think more about the 'bot' option".

CIPD's factsheet uses two of those words. Where recruitment or retention is a challenge, it says, "skills will need to be built via staff development or borrowed via outsourcing".

The Office for National Statistics (ONS) put its own question to businesses in its Business Insights and Conditions Survey (BICS): "What has been your business's approach to integrating artificial intelligence technology related skills into your workforce?" Its four substantive answers are automating or replacing roles with AI, outsourcing roles, recruiting new staff with AI-related skills, and training or retraining existing staff.

From ONS's figures for June 2026 (BICS Wave 159), the table sets two of those answers side by side by size of business; businesses could give more than one approach, and "Other", "Not sure" and "Not applicable" answers are excluded.

Size band	Training or retraining existing staff (% of businesses)	Recruiting new staff with AI-related skills (% of businesses)
0 to 9 employees	31.7	2.1
10 to 49 employees	35.6	5.4
50 to 99 employees	42.2	8
100 to 249 employees	43	7
250 or more employees	41.7	10.3

ONS says the pattern "appears to show that businesses are mainly adapting to AI through internal workforce development". In the same June 2026 figures, automating or replacing roles with AI runs from 7.6% to 12% of businesses across the bands, and outsourcing roles from 1.5% to 4.7%. With about 300 people, the firm sits in the last row.

The CIPD survey also asked about practices for reducing recruitment difficulties. Upskilling existing employees to fill hard-to-recruit-for positions is the most common, in place at 51% of organisations, with apprenticeship schemes third at 44% and increased "investment in automation (ie replacing labour with technology)" at 23%. These answers come from all 1,014 respondents (April to May 2026) and are not specific to AI.

The firm makes its choices in CIPD's terms. The vacancy's routine tasks go to the "bot" where the firm's systems allow. The redesigned HR role is offered first as an internal move, following the CIPD report's advice to treat "internal mobility as a core resourcing strategy", and whoever moves is trained for it.

For customer service and finance, the firm builds. It buys only a skill it cannot build before the skill is needed, and borrows for a short, specialist piece of work. Its decision is to build first and buy what cannot be built in time.

Testing the choices against more than one future

Before paying for anything, the firm checks whether those choices hold if AI moves faster or slower than expected. Skills England's 2026 report said of AI's uneven reach across occupations: "This unevenness complicates planning." It left its job projections unadjusted for AI because "the evidential basis for making such changes is not strong", and described them as "an indication of the scale of change in skills demand".

CIPD's third planning stage asks for this check. Scenario planning, the factsheet says, "can show how different futures might affect people requirements and help to formulate contingency and adaptive plans for achieving future goals".

The Government Office for Science (GO-Science) published AI Scenarios 2030 in June 2026 for "testing plans and assumptions against unanticipated shocks", and said its scenarios were "not predictions". Two of the five: in Slow Burn, "AI progress slows, and AI causes less disruption than expected." In Transformation Economy, "Humans are pushed 'out of the loop' for most tasks, causing widespread labour displacement".

The firm puts GO-Science's question, "if the world looked like this, what would it mean for my area?", to each re-scoped role in both futures. The report's key finding carries the test: "Even in futures with lower levels of AI capability or unemployment, the nature of work is likely to change, with routine, execution-oriented tasks increasingly becoming automated."

Whatever the firm needs in every future it tried is what it builds first. Skills England's report called such skills "no regret" skills, "capabilities that are valuable now and likely to remain essential as AI adoption grows", and said it prioritised them. CIPD's factsheet describes the direction: future roles "are likely to need greater technological and digital awareness".

How deep the training goes, and who pays

The firm now decides how deep each person's training goes and how to fund it. The Employer Skills Survey 2024, carried out from June 2024 to January 2025, counts employers by site: 59% of sites expected to need to upskill their workforce within 12 months, rising to 74% of sites using AI, against 56% of those that did not.

Skills England's 2026 report, citing the same survey, said training spend by businesses in England fell 9% between 2019 and 2024, in 2024 prices. So the firm gives the build its own line in the budget.

A January 2026 analysis of AI job adverts by Perspective Economics, for the Department for Science, Innovation and Technology and the Department for Culture, Media and Sport, sorted AI roles into highly technical experts, specialists able to implement AI techniques, such as data analysts, and broader implementers using tools such as generative AI. Its adverts dated from 2021 to 2023.

Sorted that way, most of the firm's office staff sit at implementer level and a few need specialist depth. The firm has no expert role in-house.

Existing employees can be apprentices where the job needs significant new knowledge, skills and behaviours (the apprenticeship funding rules for August 2026 to July 2027, paragraph 63).

In the CIPD survey (1,014 respondents, April to May 2026), 52% of organisations offer apprenticeships, meaning a programme is in place, rising to 72% of large organisations (250 or more permanent employees). CIPD says apprenticeships are encouraged by the Growth and Skills Levy, previously the Apprenticeship Levy, and by government support.

When the firm picks an apprenticeship standard, the 2023 report's ranking is one guide: of the standards with more than 1,000 starts, the report ranked Level 4 Business Analyst first for leading to the jobs most exposed to AI and Level 3 Data Technician fourth, a ranking built from occupational maps and 2022/23 starts data.

Building costs this firm less than hiring: its people know the business, and AI expertise is scarce, a barrier businesses report. In the CIPD survey, 16% of organisations compare the cost of developing talent in-house with recruiting. Our team training gives the many the basics; a few take AI Transformation (Level 4) or Data & AI Essentials (Level 3) apprenticeships, with project work on their own job from week one.

What the plan counts, and the people team's skills

The firm's plan also needs measures. In the CIPD survey (1,014 respondents, April to May 2026), 36% of organisations collect data in none of the nine areas it asked about, which run from skills gaps and hiring demand to the cost of investing in technology to automate. The report recommends a few "critical indicators such as skills gaps, hiring demand and retention".

The firm takes that short list and adds the cost of developing talent in-house against recruiting, the comparison it weighed when it chose to build.

The people team is in the build too. In CIPD's HR pulse report 2026, workforce planning and skills forecasting is a high or medium priority for 78% of people professionals, and 14% hold specialist skills in workforce planning. In a subset of 353 people professionals surveyed in early 2026, 75% had received no training to apply AI in their job role.

So the HR team's training goes into the plan beside everyone else's, measured with the same skills gap analysis on real work and aimed at the AI and data skills a people team uses.

CIPD calls workforce planning "an iterative process", and its April 2026 research, HR's role in reskilling employees as AI reshapes work, named the "common mistake" of treating AI skills planning "as a one-off exercise". The firm sets its review points to match: every vacancy, and each budget round.

The plan on one page

The firm's plan fits on one page, one line per decision, each with its owner.

• Trigger: The HR director reviews every leaver's role against what AI is changing in it before any advert goes out.
• The "bot": Line managers move a reviewed role's routine tasks onto the firm's systems wherever those systems allow.
• Build: The HR director offers each redesigned role inside the firm first, and line managers put customer service, finance and HR staff through training at the depth each role needs.
• Buy: The HR director recruits only for a skill the firm cannot build before it is needed, with finance approving the cost.
• Borrow: Finance approves outside help only for short, specialist pieces of work with a fixed end.
• Scenario test: The HR director and finance check every choice against a slow-change future and a fast-change one before money is committed.
• Measures: Finance tracks skills gaps, hiring demand, retention and the cost of training staff against recruiting, and reports them to the HR director.
• Review: The HR director revisits the plan at every vacancy and each budget round, with line managers reporting what AI has changed in their teams.

Frequently asked questions

How far ahead should a workforce plan look?

The next 12 to 24 months is the span CIPD's Resourcing and talent planning report 2026 recommends for scenario-based planning. In its survey (1,014 HR professionals, April to May 2026), 24% plan up to six months ahead, 23% between one and two years, and 16% over two years into the future. Large organisations tend to look further ahead than SMEs, those with up to 249 permanent employees.

What is the difference between hard and soft workforce planning?

"Hard" workforce planning, in CIPD's factsheet, "is about numbers: predicting how many people with what skills are likely to be needed". "Soft" (or strategic) planning "is about defining a strategy or developing a strategic framework within which information can be assessed". CIPD calls the whole task "as much art as science" and adds: "No single formula exists that will give a 'correct' workforce plan."

Does the people team's own strategy cover AI?

Thirty-eight per cent of people professionals in medium and large organisations agree that AI integration is a core element of their people strategy, and the same share disagree, in CIPD's HR pulse report 2026 (a base of 756, self-reported). On future skills, 61% say their people team proactively identifies the skills needed in the next one to two years.