AI Workforce Strategy Lessons from Microsoft, Barclays, and IKEA
AI Workforce Strategy: Lessons from Microsoft, Barclays and IKEA
What UK businesses can learn about automation, reskilling and workforce planning.
Three companies have put the employment question behind business AI into sharp focus. Microsoft is offering voluntary buyouts while spending heavily on AI. Barclays plans to cut 5,000 jobs as it uses AI to lower costs. IKEA automated routine customer enquiries and trained 8,500 people for more valuable advisory work. Each case is different, but they all leave leaders with the same question: what happens to the people and capacity when the technology changes?
What happened at Microsoft
On 26th October, Microsoft said it would offer voluntary buyouts to about 60,000 eligible employees across selected teams in the UK, US, Canada and India. The offer covers more than a quarter of its global workforce, but it doesn't mean there’s 60,000 confirmed job losses.
The announcement follows a separate restructuring in July, when Microsoft said it was eliminating around 4,800 roles. At the time, the company said AI wasn't replacing those roles, but it did acknowledge that the technology is changing how work gets done. Microsoft has also been redirecting very large sums towards data centres, chips and other AI infrastructure. Reuters reported that the company had forecast $175 billion in capital expenditure for the 2026 calendar year.
Headlines will inevitably compress that into 'Microsoft replaces people with AI', and they may have a very valid point. There’s no two ways about it: Microsoft is cutting costs, changing priorities, and funding a capital-intensive technology shift simultaneously, and these days, AI is massively shaping the business context. But that doesn't mean a Copilot prompt directly replaces every departing employee. And we’d go as far as to suggest that it actually never, ever, should.
Barclays brings the question closer to home
For UK businesses, Barclays brings it a little closer to home and makes the debate harder to dismiss as a purely Silicon Valley issue. On 22nd October, Reuters reported that the bank plans to cut around 5,000 jobs, roughly 5% of its workforce, while using technology (including AI) to reduce fixed costs and improve returns. Around 60% of the cuts are expected to fall in support functions, including developer roles.
The cuts arrived alongside a 12% rise in third-quarter pre-tax profit to £2.3 billion and a new £500 million share buyback. Barclays is profitable and returning capital while it changes the shape of its workforce, and while companies are entitled to make that commercial decision, leaders shouldn't hide it behind vague language about transformation.
AI can create capacity, there’s no two ways about it. Management then decides whether that capacity improves service, supports growth, reduces headcount or increases returns. Anyone with a bit of a brain can usually read between the lines and tell when that choice is being left out of the story.
What IKEA did differently
IKEA offers a more optimistic comparison that kick-started back in 2021. There, the retailer introduced Billie, an AI chatbot designed to handle routine customer questions. It then trained 8,500 call-centre employees to provide remote interior design advice and deal with more complicated customer needs.
By 2023, Billie had handled 47% of customer enquiries to call centres over its first two years, according to Reuters. In July 2026, Fortune reported that the share had risen to 74%. The same report said IKEA's remote sales centres had become its fastest-growing sales channel. Separate corporate layoffs that year didn't affect those centres and weren't attributed to AI.
So is IKEA proof that reskilling always beats redundancy? Maybe, but it’s more of a timely reminder that people buy from people, and you shouldn’t try to slash workforces just to increase margins. In fact, people can be the answer to growing them - especially in a seemingly all AI world for businesses, social media, food menus and… just about everything else, these days.
The thing with IKEA is that they had an adjacent service to grow, employees with transferable customer skills and enough scale to invest in training - a smaller accountancy firm or regional manufacturer may not have an obvious new revenue stream waiting in the next room.
The part worth borrowing, though, is how IKEA have framed the work. It started with a task that technology could handle, and asked where human judgment could create more value. The retailer automated repeatable questions and moved people towards work involving interpretation, reassurance, taste and commercial judgement. The technology changed the shape of the role rather than simply deleting the job title.
Does AI mean fewer jobs?
Honestly? It seems that every few days, a headline comes out to answer that question in a pretty resounding ‘yes’ - but only for companies that have presences across the world. If a process needs fewer hours after automation and the business can't redeploy that capacity, headcount may fall. Elsewhere, the saved time might clear a backlog, improve service, or allow the same team to support growth without recruiting at the same rate.
Most SMEs shouldn’t fear too much; instead, they face a less dramatic near-term question: what should we stop asking people to do by hand, and what should they do with the time saved instead? Many AI projects leave the second half unanswered. Saving ten hours a week sounds useful, but only if the business has decided how those hours will improve an outcome.
An AI workforce strategy links automation to the operating plan. It identifies which tasks may change, which capabilities the business must retain, how people will learn new ways of working and where human review remains necessary. It also gives the released capacity a purpose.
What can UK SMEs learn from all three companies?
Microsoft, Barclays and IKEA operate at a scale that most UK organisations probably will never resemble, but their choices still expose questions that are easier to answer before an AI rollout than after it.
Start with tasks rather than job titles
Map the repetitive work inside a role: summarising meetings, triaging requests, drafting standard replies, checking documents or moving information between systems. Jobs are bundles of tasks. AI will usually change those bundles unevenly.
Name the destination for saved time
Decide whether capacity will go towards faster response, better quality control, more customer contact, reduced overtime or growth. Without a destination, 'productivity' becomes a number in a presentation rather than an operational improvement.
Protect the knowledge behind the process
The person doing a repetitive task often understands the exceptions, weak data and awkward customer cases that are missing from the procedure. Involve them in design and testing before automating the work. Otherwise, the business may remove the very knowledge needed to supervise the system.
Train around real work
Generic AI awareness has limited value on its own. Staff need practice with approved tools, realistic scenarios, data rules and the point at which a human must step in. The UK Government's 2026 employer guide on AI upskilling reaches a similar conclusion: training works best when it's practical, linked to day-to-day tasks and built into existing workflows.
Keep governance proportionate
A low-risk drafting assistant doesn't need the same controls as a system that assesses staff performance, handles sensitive client data or makes decisions with a significant effect on a person. Define approved tools, access, data boundaries, review requirements and an owner for each use case.
Measure the whole result
Track time saved, but also look at rework, service quality, employee confidence, customer outcomes, security incidents and cost. An automation that's quick but regularly wrong hasn't removed work. It's moved it somewhere less visible.
A practical AI workforce strategy for smaller organisations
A useful first pass doesn't require a large transformation programme. Choose one workflow with enough volume to matter and a clear owner. Record how it works today, including the exceptions. Then run a controlled pilot with the people who know the process best.
- Define the business problem and the current baseline.
- Separate repeatable steps from decisions that need context or judgment.
- Check data sensitivity, permissions, retention and supplier terms before connecting AI to business systems.
- Agree on who reviews outputs and who can stop or change the workflow.
- Give the affected team protected time to test, learn and document what changes.
- Review the result after a fixed period and decide whether to expand, revise or stop.
IT, operations and HR need to work together. IT can assess the platform, data access and security. Operations can define the process and expected result. HR can support role design, consultation and training. Leave out any one of those perspectives, and the AI workforce strategy becomes fragile.
For organisations that use AI to monitor workers or to support decisions about them, the legal and ethical stakes are higher. ICO guidance says human involvement must be meaningful, with people able to check, interpret, and challenge automated recommendations rather than simply approve them out of habit. Employment and data protection advice may be needed for higher-risk uses.
Decide what the efficiency is for
Microsoft, Barclays and IKEA haven't produced three templates that other companies can simply copy. They show three possible uses for the capacity AI creates: funding a costly technology shift, lowering the cost base and increasing returns, or moving people towards a service customers will pay for.
For UK SMEs and mid-market organisations, the decision should be made alongside the technical plan. Define what the tool will do, what people will do differently and how the business will know the change has worked. Then explain that choice plainly to the people whose work will change.
A useful result is a business that handles routine work more efficiently, keeps human judgement at the heart of your business - where it matters most - and gives people a credible route into the work that comes next. Remember, you can’t have a conversation on a Monday morning with a dashboard full of automated tasks - but you can with the people who you’ve rolled AI adoption out to!
Planning AI adoption in your organisation? Fifosys can help you assess readiness, choose practical use cases and put the right Microsoft 365, data, security and governance foundations in place. Start a conversation with our team.
AI workforce strategy FAQs
What is an AI workforce strategy?
An AI workforce strategy is a plan for how automation will change tasks, skills, roles and capacity within an organisation. It connects AI adoption to business goals, employee development, governance and measurable outcomes.
Does introducing AI always mean cutting jobs?
No. AI can reduce the time needed for particular tasks, but businesses can choose to use that capacity to improve service, clear backlogs, support growth, reduce overtime or redesign roles. Employment outcomes depend on business decisions and the opportunities available for redeployment.
What can SMEs learn from IKEA's use of AI?
IKEA's approach demonstrates how automating routine customer enquiries can be combined with employee training and redeployment into more advisory, customer-facing work. Smaller organisations should assess whether similar opportunities exist within their own services and teams.
How should a small business start planning AI automation?
Start with a defined workflow, identify repeatable tasks, involve employees who understand the process, assess data and security risks, and run a controlled pilot. Measure quality and business outcomes as well as time saved.
Why is employee training important for AI adoption?
Training helps employees understand approved tools, data protection requirements, practical use cases and when human judgement is necessary. It also supports confidence and gives staff a meaningful role in improving how AI is introduced.
What should businesses measure when introducing AI?
Businesses should measure time saved, output accuracy, rework, customer service quality, employee confidence, operating costs and security outcomes. These measures help establish whether automation is creating genuine value.
Build an AI strategy around your business and your people
We help UK businesses assess AI readiness, identify practical use cases and put the right security, data and governance foundations in place. From early planning to implementation, our focus is on making technology work for the people who use it.
Talk to our team
Discuss where AI could improve your operations, support employees and deliver measurable business value.
Assess your AI readiness
Understand your current capabilities, risks and opportunities before committing to wider AI adoption.