AI at work: data surprises that are changing HR strategies.

AI’s rapid growth is fueling workforce shake-ups. For example, tech giant Meta has poured billions into AI while slashing about 8,000 jobs (10% of its staff) in a recent restructuring ([1]) – a move that reportedly left remaining employees seething with 'anger and anxiety' ([2]). Across industries, AI has been cited as the reason behind nearly 11,000 U.S. layoffs in the past month alone and over 100,000 job cuts in the first half of 2026 ([3]).

Yet these alarming headlines don’t tell the whole story. Global analysts say AI’s net impact on jobs could be positive in the longer term – the World Economic Forum projects that by 2030, AI and related technologies will create roughly 170 million new roles globally while displacing about 92 million ([4]). That’s a net gain of 78 million jobs, but with a critical caveat: the workers losing roles are not the ones filling the new jobs, resulting in a profound reskilling challenge ([5]). In fact, more than half of the global workforce – 59% – will require some form of upskilling or retraining within the next four years to keep pace with AI-driven changes ([6]).

Some shifts are already apparent. With AI automating routine tasks, entry-level opportunities in certain fields are shrinking even as demand for higher-skilled talent grows ([7]). For instance, a study found the number of employed U.S. software developers aged 22–25 – typically entry-level – has dropped nearly 20% since 2024, even while overall software developer employment rose for older workers ([8]). As one industry analyst put it, the 'entry-level ladder is being pulled up' by automation, not the entire building ([9]). This puts pressure on leaders to find new pathways for junior talent, through reskilling or redeployment, so early-career employees aren’t left stranded.

Meanwhile, other data suggest that organizations embracing AI as an augmentation tool rather than pure automation are experiencing unexpected workforce growth. A PwC analysis found that companies most able to harness AI increased their headcount by 52% since 2018, far outperforming the 36% growth at less AI-driven companies ([10]). These high performers – what PwC calls 'superstar' AI organizations – also achieved nearly five times the labor productivity gains of their peers by using AI to amplify (not replace) human expertise ([11]). And in an eye-opening twist, the wave of big-company layoffs is creating opportunities elsewhere: 81% of small and mid-sized businesses in the U.S. and U.K. report they’ve benefited by recruiting talent shed by larger firms during AI restructurings ([12]). Nearly half of those smaller companies have already hired someone laid off by a bigger employer’s AI-driven cuts ([13]), and 67% plan to continue growing their workforce through the end of 2026 to seize new opportunities enabled by AI ([14]). For large enterprises, it’s a cautionary lesson: aggressive cost-cutting in the name of AI can end up handing valuable talent – and a competitive edge – to others.

Many companies are discovering that capturing AI’s benefits requires not just new technology, but new ways of working. One major hurdle is the failure to update roles and workflows to properly integrate AI. In one recent global survey, 98% of executives said they plan to change their organizational design in the next two years for AI, and 65% expect 11–30% of their workforce will need redeployment or reskilling due to AI impacts ([1]). But so far only 15% of organizations have reached the 'optimizing' stage of AI adoption where such role redesigns actually happen ([2]). In other words, almost everyone intends to rethink jobs for AI, but very few have gotten it done.

A small minority is ahead of the curve. In Kyndryl’s 2026 People Readiness study, roughly 9% of organizations – dubbed 'AI pacesetters' – stand out for already redesigning roles around AI rather than just layering new tools onto old job descriptions ([3]) ([4]). These pacesetters were 1.5 times more likely to achieve AI-related revenue growth and significantly more likely to report product innovation gains than their peers ([5]). Their approach starts with mapping what people in each role actually do all day and deciding how those tasks change: which tasks can be handed off entirely to AI, which should stay but evolve (for example, drafting a report shifts to reviewing AI-generated drafts), what new duties (like supervising AI outputs or handling exceptions) emerge, and which human tasks should remain untouched for quality or ethical reasons ([6]). They then rewrite job definitions accordingly – a step too often skipped in traditional tech implementations.

Why isn’t everyone doing this? Simply put, true role redesign is hard. It requires significant time and attention from senior leaders, and it can be an uncomfortable process. Line managers must honestly dissect their teams’ workflows, agree on what work will change or stop, and accept that some familiar tasks may be relinquished to machines ([7]). This kind of deep organizational change is slower and more contested than just deploying a new software tool – there’s no quick demo to wow the board in a few weeks, which can make it a tough sell for impatient stakeholders ([8]). As a result, many companies race to implement AI technology first and promise to adjust roles later, but by then employees have already invented their own ad-hoc ways to incorporate the tech, making formal changes far more difficult ([9]). This 'technology-first, people-later' pattern leaves organizations stuck in pilot mode, with workers absorbing extra duties or confusion instead of achieving true productivity breakthroughs.

The cost of postponing role redesign is evident in the data. An analysis of 120,000 workers found that while 82% of employees who started using AI continue to use it over time, only about 2% have actually reached the stage of consistently integrating AI into their day-to-day workflows ([10]). The vast majority remain stuck in basic experimentation or occasional task automation, indicating that companies are rolling out tools faster than they are reengineering processes to support new ways of working ([11]). In other words, technology is racing ahead, but many organizations’ structures and practices are holding people back from using AI to its full potential.

When companies do take the time to redesign work thoughtfully, the payoffs are real. Boston Consulting Group’s latest global “AI at Work” survey revealed that employees in organizations pursuing end-to-end workflow redesign were 24 percentage points more likely to see significant performance improvements – and 20 points more likely to report increased job satisfaction – compared to those at companies that only introduced AI tools without rethinking processes ([12]). Clear strategy and role clarity can amplify AI’s impact: providing employees with well-defined AI integration plans and process changes boosted measurable business impact by 25 percentage points, whereas simply adopting new tools without such strategy improved performance by only ~5 points ([13]). The lesson is that redesigning workflows and roles unlocks far greater value than superficial automation. By treating AI as a way to augment human strengths rather than replace people, these organizations are not only seeing productivity gains but also a more engaged, adaptable workforce.

For instance, OpenAI’s chief financial officer – who rebuilt her finance team to be an 'AI-native' function – argues that AI helps her staff surface insights faster but does not replace human judgment or accountability in decision-making ([14]). She offers practical steps such as giving every employee access to AI tools and redesigning workflows around key decisions (instead of just automating tasks) as part of building a high-performing, AI-powered finance operation ([15]). The takeaway for senior leaders is to approach AI transformation as an opportunity to reimagine how work gets done. By freeing people from busywork and refocusing their roles on what humans do best – problem-solving, creative thinking, and relationship-building – organizations can achieve both efficiency and innovation without discarding their human talent.

Even with new technology and redefined roles, the human side of AI-driven change will determine success. Many employees are at once excited and anxious about AI’s impact. Gallup finds that half of U.S. workers now use AI at least occasionally in their jobs ([1]), and BCG reports that frontline employee AI adoption has surged to 74% using it at least weekly (a 23-point jump from last year) ([2]). Yet 18% of American workers – and 23% of those in organizations already using AI – believe their job is likely to be eliminated within five years due to automation ([3]). These fears of “FOBO” (fear of becoming obsolete) are very real – if unaddressed, they can erode morale and drive people to disengage or even exit.

How leaders guide this transition can make or break employee trust. Simply rolling out AI is not enough – U.S. employee engagement has stalled at around 31% through early 2026, showing no improvement despite growing AI investments ([4]). However, when organizations introduce AI with clear goals, training, and active managerial support, employees respond very differently. Workers who are given a clear plan for using AI, coupled with supportive managers, are far more engaged (53% engaged in such environments vs just 30% where guidance is lacking) ([5]). In short, people embrace AI-fueled change when leadership provides transparency, training, and reassurance; without those, confusion and skepticism can quickly undermine transformation efforts.

Bridging the skills gap is another urgent leadership challenge. Advanced AI is changing what skills are in demand faster than traditional training can keep up. 72% of workers say the skills needed for their job are shifting due to AI, but only 36% feel they’ve received adequate upskilling to meet these new demands ([6]). The World Economic Forum foresees 59% of the global workforce will require reskilling by 2030 to adapt to AI-driven changes ([7]), yet many companies have been slow to invest in broad-based retraining. Some are now taking action: 70% of small businesses report they are investing in AI training for employees, and 64% are focusing on upskilling their current staff to address AI-related skill gaps ([8]). A number of major firms are also rethinking talent pipelines – for example, IBM’s HR chief announced plans to triple U.S. entry-level hires for 2026 and rewrite junior job roles to emphasize human judgment and oversight as AI takes over routine tasks ([9]). These moves recognize that without significant investment in people’s capabilities, even the best AI tools will fall flat.

Proactive communication and sensible policy can likewise ease the cultural strain of rapid automation. Labor unions have started to push for AI safeguards in the workplace – a recent analysis found 85 to 90 union contracts (from media to tech industries) now have explicit rules around AI’s use ([10]). These provisions require that employers give advance notice when implementing AI systems, set clear boundaries on where AI can and cannot replace human decision-making (especially in areas like hiring or discipline), provide training or reassignments for workers impacted by AI-driven changes, and establish channels for employees to appeal or correct AI-driven outcomes ([11]) ([12]). Rather than waiting for a crisis, some forward-looking HR leaders are borrowing these ideas to craft internal AI policies that include employee input and transparency from the start ([13]). By involving staff in setting guardrails and expectations for AI, companies can lower resistance, build trust, and speed up adoption. In an environment where the only constant is change, such an inclusive approach can make the difference between an AI revolution that energizes your workforce and one that leaves people feeling alienated.