In 46% of Companies, the CHRO Isn't in the Room
IBM's 2026 CHRO study quantifies what separates companies that turn AI into growth from those that turn it into cost-cutting.
In brief
IBM surveyed 1,500 CHROs and 8,800 employees in over 20 countries to figure out who actually captures the value of AI. The verdict: it's not the companies deploying the most AI, but those explicitly redrawing the line between human and machine decisions. Yet nearly half of CHROs aren't consulted when AI strategy is decided.
🍺 Bar-stool version
We spent three years wondering whether AI would replace people, and IBM shows up with a more awkward question: what do you actually do with the freed-up hours? Spoiler: in 58% of companies, nobody planned for it. They either get swallowed by more work or turned into savings, which is to strategy what eating standing up is to fine dining. Meanwhile, 60% of employees say their skills are quietly eroding — critical thinking topping the list. 🙂
Key takeaways
- 1
46% of organizations don't involve the CHRO when AI strategy is set, and 73% of CHROs struggle to align executive committee decisions with AI's impact on work.
- 2
Where the CHRO co-owns defining what stays human, 76% of organizations say employees feel free to challenge or override an AI recommendation, versus 43% when HR is only consulted.
- 3
29% of workflows already integrate AI agents, heading to 40% by 2028, but only 26% of companies clearly define what is human-led, AI-assisted, or AI-executed.
- 4
Skill erosion is the blind spot: 80% have a reskilling roadmap, but 60% of employees report a loss, with critical thinking leading the list.
- 5
Only 23% of organizations are increasing junior hiring, and just 7% cite entry-level talent as a segment being strengthened — the traditional path to expertise is closing.
- 6
Only 42% mainly reinvest AI productivity gains into innovation or training; 49% of leaders cite loss of employee trust as their top concern.
- 7
72% of organizations make limited or no use of AI within the HR function itself — the function meant to steer the change hasn't yet experienced it.
The thesis: the "Thinking Organization"
The IBM Institute for Business Value defines the "Thinking Organization" as a company that combines humans and AI to make better decisions, continuously adapts how work gets done, and converts freed-up capacity into growth. Ten capabilities, grouped into five pillars, make up a maturity index.
The ranking is brutal in its simplicity: the top quartile, 375 of the 1,500 organizations, shows nearly 2x higher revenue growth, 2x better workforce outcomes, and 2x more innovation brought to market than the bottom quartile.
An important methodological point: the level of AI adoption significantly predicts the score, but sector, geography, and size do not. In other words, this isn't about resources or industry — it's about management decisions.
IBM also honestly notes these are correlations from cross-sectional data, not proof of causation. Worth keeping in mind once the "2x" figures start circulating on LinkedIn.
The mandate nobody wants
The report's most striking finding: in nearly half of companies, the CHRO isn't in the room when AI strategy is decided. "Enterprise intelligence" is scattered across IT, operations, and finance, and only 28% of leaders say they have a shared roadmap with IT on a common cadence.
This vacuum has measurable consequences on the ground. Where the CHRO shares responsibility for defining what stays human, employee trust in AI-assisted decisions improves in 63% of cases, versus 19% elsewhere. The gap is too wide to blame on technology alone.
The report quotes Niharika Mohan Vohra (Schneider Electric, Greater India): AI is "a copilot, not an autopilot." Translated into a metric: employees supervise the machine instead of submitting to it.
And the mandate won't stay vacant: 77% of CEOs believe talent leadership and technology leadership roles are converging. If HR doesn't claim it, the CTO, CFO, or COO will.
Redesigning work, not org charts
29% of workflows already integrate AI agents, on track for 40% by 2028. But design is lagging: only 34% of roles are redesigned to emphasize judgment and exception management, 23% actively track workflow outcomes with authority to escalate or halt, and only 11% actually embed decision-making authority into execution.
The cost of this ambiguity is quantified. 36% of leaders say lack of clarity on responsibilities complicates deployment, and 43% of employees believe that when AI gets it wrong, it's them who bear the consequences.
Conversely, organizations that clearly define what is human-led, AI-assisted, or AI-executed achieve 18% risk reduction and 20% quality improvement. Those that continuously manage workforce recomposition post 4% revenue growth versus 2%. Only a quarter do this.
Anne-Catherine Ropers, Group CHRO of Crédit Agricole S.A., sums up the shift: "We don't look at jobs. We look at the process, the activity, and the task."
The blind spot: cognitive atrophy
This is the most interesting part of the report. 80% of companies have a reskilling roadmap, which addresses missing skills. None address existing skills that are degrading because they're no longer exercised. "A skills gap gets funded. Erosion remains unmanaged."
46% of leaders flag this as a major concern, 60% of employees notice it, and three out of four affected employees say AI has already begun eroding some of their skills. The one cited most: critical thinking — precisely the one 57% of leaders deem essential.
The perception gap is striking when it comes to evaluating AI outputs: 71% of leaders prioritize the ability to supervise, validate, or override AI, versus 38% of employees. On exception management, it's 32% versus 14%.
And the learning pathway is closing: only 23% of organizations are increasing junior hiring, and when asked which segment is being strengthened to support future capacity, entry-level talent gets just 7%, far behind experienced contributors (34%) and frontline managers (27%).
Where the gains go, and why HR must go first
"AI deployments always end up at the same question: where do the gains go?" Only 42% of organizations mainly redirect them toward innovation or training. The rest either arbitrate, pocket the savings, or simply have no pathway from productivity to opportunity.
The report also flags invisible labor: 80% of leaders acknowledge that AI adoption creates unrecognized tasks — validating, correcting, contextualizing, handling exceptions — and 42% of employees say AI increases their workload or leaves this work in the shadows.
Another surprise: leaders underestimate their teams. 82% of employees feel free to challenge a flawed AI output, while leaders credit them with only 59%. On job security sentiment, it's 81% versus 55%.
Hence the fifth, most uncomfortable pillar: 72% of organizations make limited or no use of AI within HR itself, and only 13% of leaders rate their HR team's AI literacy as high. The function meant to steer the transformation hasn't yet practiced it.
IBM's own experience, mistakes included
The most concrete part of the report comes from Nickle LaMoreaux, IBM's CHRO, on the AskHR program launched following the 2016 rollout. First lesson: they built the tool, and nobody used it. Employees kept calling, emailing, and dropping by their HR contact.
The radical decision: cut off the HR phone line and email address overnight, worldwide. Employee NPS dropped from +19 to -35, a 54-point loss. The example is openly presented as not reproducible as-is.
The current tally: 16.3 million transactions per year, 95% of questions resolved without human intervention, CSAT at +55, and HR operating budget cut by 40% over four years.
The lesson learned fits in one sentence: AI transformation isn't a tech project with a change-management add-on — it's a behavior-change project that technology makes possible.
“« Pour la première fois en 80 ans, les DRH pourraient diriger au lieu de faciliter ou de soutenir. » — Janine Vos, Rabobank”
“« La maîtrise sans le jugement aide simplement une organisation à se tromper plus vite. » — Dr. Amit Das, Bennett Coleman & Co.”
“« Si le réinvestissement, c'est dix minutes gagnées par personne et par jour pour envoyer un e-mail de plus, ce n'est pas un bon usage du temps de qui que ce soit. » — geneviève bich, METRO”
Why it matters
This report matters beyond the HR circle because it shifts the question: it's no longer about how many jobs AI eliminates, but who decides what humans keep, and where the freed-up hours go. The most useful numbers here aren't the "2x" figures — correlations from self-reported data, which IBM honestly avoids presenting as causation — but the actual practice rates: 11% decision authority embedded in workflows, 25% continuous recomposition, 42% reinvestment of gains. The governance layer everyone talks about barely exists in practice. One bias to keep in mind: IBM sells consulting and enterprise AI platforms, and the report naturally steers toward a need for their help. But two signals deserve to be taken seriously regardless of the vendor: the erosion of critical thinking reported by employees, and the quiet closing of junior positions through which expertise used to be built. Protecting this year's productivity by mortgaging next decade's talent pipeline is a trade-off nobody explicitly owns up to — and this report at least has the merit of putting a number on it.
For teams
The same machine, on your topics.
Your sources, your team, every morning, in your language. Pilot open to three companies.
Read next
#regulationTodayCalifornia Reins In Data Centers: Seven Laws at Once
Mandatory reporting on water and electricity, and no more environmental fast-tracking: Sacramento wants the AI industry to pay its own grid bills.
Source · Governor of California (gov.ca.gov) · Governor Newsom signs most comprehensive data center laws in the nation, providing communities more control on water, electricity, and land use
McKinsey: 13 Frontier Technologies, and the AI That Absorbs Everything
The fifth edition of the Technology Trends Outlook puts $223 billion on energy, $124 billion on AI — and reveals a far less discussed signal on the jobs front.
Source · McKinsey & Company · McKinsey Technology Trends Outlook 2025