Introduction
AI is officially a mainstream workforce strategy. No longer a specialist concern, it is reshaping how organizations define performance, make talent decisions, and connect workforce capability to business outcomes.
For HR and business leaders, the question has shifted from “Does AI matter?” to “Are our people equipped to use it well?”
This shift has created a new kind of operating challenge. When we treated AI as a localized experiment, organizations could rely on informal judgment. Today, AI is a daily reality for most teams, and leaders need a reliable, shared language that sets the foundation for real decisions: how to clearly define, recruit, measure, reward, and grow AI skills.
The consensus gap
HiBob’s research shows that the market isn’t suffering from a lack of effort.
It’s not that organizations are stalled on developing AI skills. Everyone’s doing something, but there’s no consensus on a “right” approach to training and development or what to offer their people.
As it stands, the majority of organizations (73%) say they invest in upskilling, but the specific supports remain thinly distributed and inconsistent.
Defining the behaviors that drive value
To get past the guesswork, we asked 1,200 AI decision-makers questions to pinpoint the exact AI behaviors that drive real business value.
By defining a core set of observable AI behaviors, employers can:
- Source talent more accurately
- Develop their people more effectively
- Avoid relying on vague signals that can reinforce inequity
Methodology
This report is based on a survey conducted by HiBob with fieldwork support from Censuswide.
The study included 1,200 AI decision-makers across the United States, the United Kingdom, DACH, the Nordics, Benelux, and Australia/New Zealand, with 200 respondents per market grouping. Fieldwork ran from February 3 to March 12, 2026.
Respondents:
- Worked in organizations with 50 to 50,000 employees
- Held middle-management-level responsibility or above
- Had personal involvement in AI-related decisions over the previous 12 months
This data provided the foundation for our AI skills framework and assessment tool—a practical response to the need for disciplined, governed execution.
Executive summary
AI is changing work faster than traditional job titles or degrees can keep up. While 75% of decision-makers expect moderate AI proficiency to be standard across most roles within 24 months, many organizations are still stuck in a “throw spaghetti at the wall” phase of experimentation.
This fragmented approach has created a significant operational gap. Organizations are already linking AI proficiency to promotions (67%) and performance ratings (50%), but they’re doing so without clear, shared standards.
Today’s professionals need clear guidance on what behaviors earn these rewards, and companies need consistent standards to evaluate them fairly.
Key findings
- AI is the new baseline: 75% of decision-makers expect moderate AI proficiency to become standard across most non-technical roles within the next 24 months.
- AI is a gatekeeper for growth: Organizations are actively weaving AI capabilities into their formal talent decisions—67% of respondents say their organizations link AI skills to promotion criteria, and 50% tie them directly to performance ratings.
- High intent, fragmented action: While 68% of AI decision-makers claim their organizations have a defined AI sourcing strategy and 73% say they invest in upskilling, execution is piecemeal. Only 24% say they use specific levers, like ATS tagging, and no single training topic exceeds a 27% adoption rate.
- The manager readiness bottleneck: Direct managers are the top group expected to build AI capability across their teams, but responsibility doesn’t equal readiness. Only 36% of respondents see them as highly prepared to upskill their people.
- Quality control beats coding: AI decision-makers care about safe, dependable execution over flashy technical tricks. Proactively reviewing output quality (52%) and documenting workflow decisions (52%) rank as the most critical everyday AI behaviors.
- Premium pay for scarce skills: The hardest-to-recruit skills command the highest salaries. Employers are willing to pay at least a 10% premium for expertise in automation (34%), AI safety and governance (34%), and output evaluation (33%).
Recommended For Further Reading
AI success depends on building an operating model, not a tool stack
True AI transformation is an operating model challenge, not a tool adoption race. To move beyond vague buzzwords like “AI literacy,” organizations must establish a behavioral framework that defines exactly what “good” looks like.
When organizations adopt a shared language for AI skills, they can:
- Source talent more accurately by moving beyond generic hiring signals
- Develop people effectively through training tied to day-to-day tasks
- Ensure fairness by removing the “vague signals” that often reinforce inequity
This report provides the foundational data—validated by 1,200 global decision-makers—that underpins our AI skills framework and assessment tool, along with the operating context they’re designed for. Used together, the framework and tools give organizations a clear blueprint for building a scalable workforce infrastructure fueled by AI and driven by humans.
AI skills are now a mainstream business priority
AI capability is no longer a fringe technical specialty. It’s actively reshaping everyday work.
In fact, 75% of decision-makers expect moderate AI proficiency to become the standard for most non-technical roles at their organizations by 2028.
No longer considered a niche skill, AI proficiency is already a gatekeeper for career growth, and organizations are weaving AI capabilities into their formal talent decisions rather than treating them as a novelty.
Right now, 67% of companies link AI skills to promotion criteria, and 50% tie them directly to performance ratings.
How do AI expectations change by market?
While this development is undeniably global, it plays out slightly differently across different markets.
- The Nordics lead the charge, with 79% of decision-makers expecting AI proficiency to become standard at their organizations
- The United Kingdom, DACH, and Benelux closely follow at 77%
- The United States sits right around the global average at 74%
- Australia and New Zealand (ANZ) report a slightly lower expectation at 68%
The overall signal is clear: AI skills are going mainstream everywhere.
It’s time to stop building AI talent piecemeal
Now that AI skills are officially tied to performance and promotions, it can be easy to assume organizations have a watertight plan to source and develop them. This is far from the truth.
If there’s one theme that connects all of our data from this study, it’s the massive gap between intent and infrastructure. Companies want AI talent, but they’ve been trying to build it piecemeal.
There’s a sourcing gap: Intent doesn’t match action
A solid 68% of decision-makers say their organization has a defined strategy to find AI-skilled candidates. That high-level confidence masks shallow implementations of any strategy. For example, only 24% use specific levers like applicant tracking systems (ATS) tagging or dedicated talent communities.
To complicate matters, companies are missing out on the talent they already have. Only 23% of respondents say their organizations design their internal mobility and reskilling pipelines specifically to source AI-skilled employees.
The L&D paradox: High investment, no consensus
The exact same pattern plagues L&D. While 73% of respondents say their organizations invest in AI upskilling, there is no consensus on what actually works. Instead of a cohesive program, companies offer a fragmented mix of tools and upskilling supports.
Fewer than a third of respondents report that their organizations use even the most common programs—like funded learning (30%), prompt and procedure libraries (29%), and protected practice time (29%).
Likewise, our data shows that organizations are taking a highly fragmented approach to training content, with no single topic exceeding a 27% adoption rate. However, the specific topics companies are training on reveal a clear pattern: Organizations are moving away from general “inspiration and innovative productivity” and are instead prioritizing work design, control, and judgment.
The most common training topics reflect businesses’ need for safety and reliability:
- Workflow redesign (27%)
- Security and acceptable use (26%)
- Documentation (25%)
- Data reasoning and verification (24%)
Organizations know they have to develop “something” concrete to train their workforces and demonstrate being “AI-forward,” but AI upskilling remains exploratory and un-systematized. Without connecting these fragmented efforts into a reliable system, companies leave actual skill development up to chance.
Give managers the tools they need to stop improvising
Organizations are piecing together fragmented training programs and hoping for the best. But who do they actually expect to make sure these new skills stick?
The answer our data reveals is one of the biggest operational bottlenecks businesses face today.
Responsibility doesn’t equal readiness
Globally, direct managers and team leaders are the number one group organizations expect to take responsibility for building AI capability across their teams (24%). But responsibility does not equal readiness.
Among the respondents expecting managers to take on this critical responsibility, only 36% see them as highly prepared to upskill their teams proficiently in AI. This means that the majority of organizations today are asking managers to lead a massive workforce transformation, despite low confidence in their ability to do that work.
Helping managers build real AI proficiency will take more than intent. It will take structure, shared definitions, the right tools, and the training to use them well.
The “who owns it” question doesn’t have a universal answer
While managers bear the highest expectations overall, organizations are placing the burden of AI upskilling on different shoulders in different markets.
- The US: Technology providers (28%) slightly surpass managers (24%) as the primary driver of AI capability
- Benelux: Employers at the corporate level (25%) are expected to take the reins rather than individual managers (13%)
- The Nordics: The strategy here moves beyond the workplace entirely and relies heavily on vocational education and training pathways (25%) to build these skills
No matter your region, the underlying truth is the same: We cannot ask people to improvise AI transformation.
When we expect managers (or anyone else) to evaluate and coach AI skills without a shared standard, we get inconsistent, fragmented results.
The solution is to provide leaders with a concrete AI skills framework and assessment tool. Give them the exact behavioral rubrics they need to align their coaching, standardize their hiring, and assess performance fairly.
What 1,200 leaders value most
Once you give managers a behavioral framework to evaluate their teams, what exactly should they be looking for?
When it comes to AI, businesses care less about the “wow” factor and much more about safe, dependable execution. The most important AI behaviors aren’t flashy technical tricks. They are judgment, reliability, and governance.
Based on peer feedback from our study, the most critical everyday AI behaviors are proactively reviewing output quality (52%) and documenting workforce decisions for reliability (52%).
Scarcity of skills tells a different story
The skills that keep AI safe and scalable are the skills organizations struggle to find.
When we asked AI decision-makers which capabilities they find hardest to recruit for, they pointed straight to risk management and process design.
The top three hardest-to-find skills map directly to those priorities:
- AI safety
- Ethics and governance
- Workflow evaluation
Because these skills are so rare, organizations are willing to pay premium compensation rates for them.
We found that respondents indicated their organizations are willing to pay salaries at least 10% above the baseline for people with expertise in automation and technical integration (34%), AI safety and governance (34%), and output evaluation (33%).
This shows that specialized, scarce skills naturally drive up salaries, even when they are only needed in a few specific roles.
The professionals who master this—the ones who bring deep technical integration, strong governance, and rigorous quality control to the table—are leading the AI transition, and they will command the premium salaries to prove it.
Source top AI talent by defining premium skills
Knowing which skills command a premium is only half the challenge. Vague concepts like “AI literacy” won’t help a manager or recruiter spot a candidate with strong governance instincts, and they certainly won’t help justify a 10% salary premium.
To capture this high-value talent, your leaders need to know exactly what to look for. This means moving beyond the buzzwords and defining the exact behaviors that matter.
Conclusion: Build a lasting AI upskilling infrastructure
To master AI skills sourcing and building, organizations need a shared AI skills framework.
As the data shows, we cannot simply buy our way into AI transformation. We have to build it from the ground up.
To move beyond fragmented training and haphazard experimentation, we must embed a behavioral framework directly into the systems that already shape our workforces: Our job architecture, compensation bands, and learning and development pathways.
Winning organizations don’t just hand out software licenses and hope for the best. They govern AI adoption through a shared language of skills and complementary behaviors.
Establishing a behavioral standard removes the guesswork. It gives managers the clarity they need to evaluate performance fairly, it gives recruiters the exact signals they need to identify scarce talent, and it gives each team member a clear roadmap to advance their careers.
Now is the time to stop treating AI as a separate, localized experiment and start treating it as a core operating discipline as we head into the future.
