Authored by Barry Sacks, Group CTO at La Fosse
Nobody seems able to agree on what AI is going to do to the workforce.
Deloitte says more than half of UK finance chiefs (52%) expect AI to cut their headcount within a year.1 So far, though, fewer than one in ten businesses using AI say it has reduced their headcount, according to the ONS.
Graduate job adverts are down 43% in a year, according to Adzuna,3 and ONS figures show one in six young people in the jobs market are out of work.4 Yet employers told ManpowerGroup that AI skills are now the hardest to hire for.
It sounds contradictory, because it is – and I don’t think that’s going away any time soon. Even the most senior technology and business leaders still disagree on whether AI will remove large numbers of roles, or simply change their shape.
The actual problem is that businesses still have decisions to make while the debate carries on. Whether to keep investing in juniors, or stop recruiting them altogether. Whether to let teams use AI freely, or hold back. Whether introducing AI agents is making the business more productive, or just adding a new layer of problems. There is confident advice and data, pointing in every direction.
Under that pressure, many leaders are freezing, waiting for clarity that may never come. Others are forcing AI into workforce planning built for an older model of work. And some are rushing to replace people with technology before they understand what work is changing, and what capability they need instead.
Those are not strategies. They’re what happens when leaders are working with blind spots they don’t know they have. I speak to a lot of organisations about this, and the same three keep coming up.
Planning Around Headcount When the Ground Has Already Moved
Annual headcount planning assumes a degree of stability. AI doesn’t move at that pace. Its capabilities can change within a quarter, so a plan can be perfectly sensible when it’s written, and out of date by the time it’s approved.
Businesses are planning – that’s not the blind spot. The blind spot is planning around headcount at all. Headcount tells you how many people you have. Capacity tells you what the business can actually deliver, and as AI takes on more of the work, the two are drifting apart. So plan for capacity, not headcount, and review it quarterly rather than annually. You don’t need to know what AI will be capable of in two years to make a better decision about next quarter.
Hiring for Roles That AI Is Quietly Dismantling
Job titles and descriptions have a habit of sticking around after the work underneath has changed. Someone leaves, the old description comes out, it gets tweaked around the edges and recruitment starts. Meanwhile, AI may already be doing part of that job.
Start one step back. As AI agents take on more of the work, the team itself may need a different shape, not just the roles inside it. Then, before opening a vacancy, ask what outcome you need that person to deliver now, not what the previous person did. If the answer has changed, the job spec should too.
Once the role is filled, treat the job description as a starting point. Revisit it every few months, and build in time for that person to keep learning as the work shifts.
Believing Reskilling Is On Track When Delivery Has Fallen Behind
Most businesses are happy to talk about reskilling. Budgets are allocated and programmes are being built – supported by real desire and intention. But turning that intention into capability quickly enough is the hard part, and many businesses don’t realise they’re falling behind.
There’s a timing problem too. The return on reskilling can take years to show, while budgets, targets and leadership priorities still tend to reset every 12 months. So reskilling loses out to whatever pays back faster. Checking progress every quarter makes sense, but expecting the full return within a single budget year doesn’t. Treat training like any other investment, and go back months later to ask whether people are using what they learned and whether the team can deliver more as a result. A completed course tells you training happened, not that reskilling did.
Start With What You Can Control
None of these suggestions answer the big question about what AI will ultimately mean for jobs – and that’s the point. The technology, the labour market and the evidence are all still moving. Waiting for a definitive answer is still a decision, and rarely a good one.
In my view, the more useful approach is to focus on the decisions you can improve now. Review capacity more often. Define vacancies around the outcomes the business needs, not the roles it has always employed. Measure reskilling by whether it changes what your teams can deliver, not whether training was delivered.
At La Fosse, we’ve established an AI Workforce Council: ten senior leaders across technology, data, cyber, retail, media, procurement and consulting who look at how organisations can plan, resource and reskill their workforces as AI develops in real time.
Those conversations have shaped our new, free Beyond Headcount blueprint, a practical roadmap for businesses planning towards 2027.
About La Fosse
La Fosse partners with ambitious organisations to build high-performing, agile teams – connecting the right people with the right opportunities from the classroom to the boardroom. Since 2007, La Fosse has helped businesses scale and lead digital transformation through authentic relationships, bold thinking, and proven expertise.
It offers four specialised services:
- La Fosse Recruitment: High-performing talent across all levels, from technical experts to senior leaders.
- La Fosse Executive: Exceptional technology leadership via executive search and interim appointments.
- La Fosse Academy: Training and launching the next generation of diverse tech talent.
- Inovus: A challenger consultancy delivering practical solutions to data, AI and transformation challenges.
As an employee-owned company, La Fosse’s approach has also earned national recognition, including:
- Top 100 Recruitment Agencies in the UK – 6th Place (2025)
- Best Academy/Training Provider – Women in Tech Employer Awards (2025)
- Tech Recruiter of the Year – UK Business Tech Awards (2025)
- Best Large Recruitment Business – Global Recruiter Awards (2025)
La Fosse envisions a world where businesses and talent grow together – where careers are built with confidence, and people remain at the heart of every transformation.
