Every AI readiness stat you've seen this year is an average. And averages are exactly what’s getting CPOs into trouble right now.
McKinsey reports that 88% of companies are using AI in at least one function, while a separate McKinsey study found only 1% of leaders call their organization mature in AI deployment. That gap gets quoted constantly, and it's real. But it's also a company-level number, and companies don't adopt AI. Functions do. Roles do. Individuals do, at wildly different speeds.
HiBob's 2026 research adds another layer to this. 75% of decision-makers expect moderate AI proficiency to become standard for most non-technical roles by 2028, and 67% already tie AI skills to promotion criteria. When AI proficiency starts showing up in how individual performance and progression get evaluated, an organization-wide readiness average simply isn't granular enough to act on.
I've sat in enough of these conversations now to know the pattern. A CPO tells me their org is “70% AI-ready.” I ask what that means for finance versus customer support versus HR itself, and the room gets quiet. Nobody actually knows. They know the company number. They don't know where the number is hiding a 90% gap in one function and a 20% gap in another.
That's the skills gap you can't see. Not the one on the dashboard. The one underneath it.
The short version: a company-wide AI readiness score tells you almost nothing useful, because AI readiness varies enormously by role and function within the same organization. A single number can hide a function that's 90% behind sitting right next to one that's already ahead. Real workforce planning happens at the role and function level, not the company level.
Think about what an org-wide readiness score does to decision-making. It tells your leadership team “we're mostly fine,” so training budget gets spread evenly across departments. Tools get rolled out the same way to everyone. Change management gets one plan instead of ten.
Meanwhile, the World Economic Forum projects 39% of core skills will change by 2030, but that shift isn't evenly distributed either. Some roles are being rebuilt from the ground up. Others barely move. If you fund training and redesign as if every function faces the same 39%, you're overinvesting in the roles that don't need it and starving the ones that do.
The workforce itself reflects the same unevenness. HiBob research found that 44% of professionals worry AI could take their jobs, while almost exactly the same share, 45%, believe AI could make them more productive. AI readiness isn't just about adoption; it's about understanding where confidence, capability, and support diverge across your workforce.
I've said this before, and I'll keep saying it: AI readiness isn't a technology question; it's a workforce planning question. And workforce planning has never worked at the company level. It works at the role level, the function level, the individual level, which is exactly the case we made in Workforce Planning in an AI-First Environment. Aggregate readiness scores are comforting. They're also close to useless for deciding what to actually do next.
When you can see readiness by function, three things happen that a company-wide number can't give you.
You stop guessing where to spend. If legal is at 30% readiness and marketing is at 75%, that's not a coincidence you average away. That's a resourcing decision. Different budget, different timeline, different vendor, maybe a different leader entirely.
You catch the manager gap before it becomes a rollout problem. Managers are the hinge point for every AI initiative I've watched succeed or stall. A function-level view tells you exactly which manager layer is behind, instead of discovering it three months into a rollout when adoption quietly flatlines.
You can actually defend the roadmap. “We're investing here first because this function is furthest behind and highest risk” is a sentence a board wants to hear. “We think we're about 70% ready overall” is not. This is the same logic behind our Org Design and Workforce Planning practice: you redesign the roles that have actually changed, not the whole chart at once.
This is the problem livingHR's AI-People Readiness Index and HiBob's Skills module were each built to solve from opposite ends. APRI tells you where your organization stands on the things that predict AI success: governance, leadership alignment, change capacity, data infrastructure. HiBob's Skills module tells you, in real time, what your people can actually do right now, role by role.
Put those two together, and you stop getting a company score. You get a diagnostic that shows you exactly where structure and skill are misaligned, function by function, before you commit a roadmap to it. It's the kind of work we do inside our AI-Powered People Solutions practice every day.
We're working with HiBob on the full co-branded benchmark, and it's going to get specific: readiness gaps mapped by role and function, not just by company. For now, the takeaway is simpler than the report will be.
If someone hands you a single AI readiness number for your entire organization, ask them to break it apart. What you find underneath it is usually the whole story. The AI-People Readiness Index is where that breakdown starts. [Take the Index.]
Why is company-wide AI readiness a misleading metric?
Because AI adoption doesn't happen at the company level; it happens role by role and function by function. A single average can hide a function that's far behind sitting next to one that's already ahead, which leads to training budget and rollout plans that are wrong for almost everyone they touch.
How do you measure AI readiness by role or function?
You need two things: a structural readiness benchmark (governance, leadership alignment, change capacity, data infrastructure) and a real-time skills inventory of what people in each role can actually do. livingHR's AI-People Readiness Index covers the first; HiBob's Skills module covers the second. Together they show where structure and skill are misaligned by function, not just company-wide.
What's the difference between an AI skills gap and an AI readiness gap?
A skills gap is about capability: what a person or team can currently do with AI tools. A readiness gap is broader: it includes governance, leadership alignment, and change capacity alongside skills. A role can have strong skills but low readiness if the organization around it isn't set up to support how those skills get used.
None of this means throw out your company-wide number. It means stop treating it as the answer instead of the starting point. The real work, the work that actually changes outcomes, happens once you break that number apart by role and function and get honest about what you find. Some teams will be further along than you thought. Others will need more attention than a single score ever let on. Either way, you can't fix what the average is hiding. You can only fix what you're willing to look at directly.