People leads and HR directors at SaaS companies between $1M to $50M ARR whose job descriptions and hiring criteria still treat AI fluency as optional while the work already assumes it daily.
Pull any job description from your current open roles. Count how many of them name AI fluency as a requirement. Now count how many of those roles already assume it in the day-to-day work.
The gap between those two numbers is the HR problem most organizations have not named yet. The work changed. The criteria did not.
Every hire made against the old criteria arrives needing reskilling from day one. That cost is invisible until it lands in the performance review.
Over 1.5 million people in the UK received free AI training in the past year through government and Microsoft programs. 1 They received it because employers were filtering for AI fluency in interviews before HR processes ever documented it as a requirement.
At London Tech Week, the conversation confirmed it. The skills gap is not a training supply problem. It is a process design problem.
The hiring criteria, the performance measures, and the role definitions have not caught up to what the work now requires.
The people teams that compound are not the ones that ran the most training sessions. They are the ones that redesigned how the organization hires, defines roles, and measures performance.
Relve rates this 72/100, a high signal for people leads whose hiring criteria and reskilling plans have not caught up to how the work has changed.
The hiring criteria gap compounds across every function. See where the same shift is breaking the loop in other teams.
Your Job Descriptions Describe a Role That No Longer Exists
Most job descriptions were written before AI changed the role. They describe the skills and performance expectations of a version of the job that existed 18 months ago.
The person hired against those criteria arrives into a role where AI fluency is assumed in every workflow, every deliverable, and every team interaction. The criteria said nothing about it.
This is not a recruiter problem. It is a leadership failure in process design. The organization failed to update the definition of the role when the work changed.
The candidate fulfilled the contract exactly as written. The contract was wrong. That is a process failure, not a hiring failure.
The performance gap that follows is predictable. A hire without AI fluency in a role that assumes it does not underperform on the stated criteria. They underperform on the unstated ones.
The manager sees the gap but cannot articulate it because the job description does not name what is missing. Performance reviews become vague. Feedback becomes circular.
Does AI fluency actually predict performance in this role? If the work already assumes it, the answer is yes. The hiring criteria should name it. The performance review should measure it.
Most organizations have named none of these three requirements in any job description written in the last two years.
What naming it actually looks like matters. Not "uses AI tools effectively." That is vague enough to mean anything.
For a content role: "produces first drafts with AI assistance and identifies where AI output requires human judgment." For an ops role: "maintains the data consistency standard across three functions using AI-assisted monitoring."
A criterion that specific is evaluable in an interview and measurable in a review. A criterion that vague is neither.
The gap is not closing on its own. Every quarter the work assumes more AI fluency. Every quarter the criteria stay the same.
Most Companies Treat Reskilling Speed as a Benefits Line
The organizations that close the AI fluency gap fastest keep their people through every shift. The ones that do not hire from a shrinking pool of candidates who have already closed it elsewhere.
The NHS moved 500,000 staff onto Copilot. That pace was not possible because they offered optional training. 2
It was possible because someone redesigned the workflows, the role expectations, and the support structures before the tool went live.
The reskilling was built into how the work was done, not added on top as a benefit. Most SaaS companies are doing the opposite.
They offer an AI tools budget. They run a lunch-and-learn. They add an AI fluency paragraph to the company handbook.
None of those actions close the gap because none of them change what gets measured in a performance review or required in a job description.
A reskilling program that runs once is a campaign. An organization that builds AI fluency into the role definition, the hiring criteria, and the performance measure creates a compounding loop.
Every new hire enters at the new baseline. Every existing employee is measured against it. The baseline rises with every iteration. That is the competitive difference between a campaign and a loop.
- At least one open role has been rewritten in the last 90 days to name AI fluency as a requirement
- Your most recent performance review cycle measured AI fluency in at least one role where the work assumed it
- Reskilling outcomes are tracked as a team metric not just as individual training completion
- You know which roles carry the highest AI exposure in your organization
- You have a documented fairness rationale for how reskilling resources are allocated across those roles
Three or more unchecked means the organization is running campaigns, not building a loop. The gap between where the work is and where the criteria are will keep widening.
The Equity Problem Is Inside the Gap Not Separate From It
AI exposure is not evenly distributed across roles. Administrative and coordination roles carry the highest exposure. Those roles are disproportionately held by women and by people earlier in their careers.
When an organization closes the AI fluency gap without an equity plan, it closes it unevenly. The people with the highest exposure and the least access to reskilling resources fall furthest behind.
This is not a diversity initiative. It is a legal and operational risk, and most HR teams are not thinking about it yet.
A performance review that penalizes employees for not having AI fluency in a role that was never redesigned to require it carries a disparate impact problem embedded in it.
Most HR teams are not thinking about this yet. They will be after the first employment dispute, when the process gap becomes a legal one.
An equity plan built into the role redesign process creates clarity. An afterthought added to the company handbook creates ambiguity that surfaces in exactly the wrong moment.
The equity answer is not complicated. Map the roles with the highest AI exposure. Allocate reskilling resources to those roles first.
Document the rationale before the redesign goes live. Three steps, one document, done before the first revised job description is posted.
Redesigning roles for AI fluency without a documented fairness rationale creates legal exposure. Build the equity answer into the process design before the first revised criteria go live, not after the first dispute.
Three HR Decisions Before the Next Hire Goes Live
Pull five job descriptions for roles where the work already assumes AI fluency daily. Count how many of them name AI fluency as a requirement. Count how many measure it in the performance criteria.
The gap between those numbers is your starting point. You do not need to fix all five today. You need to know the number.
Owner: People lead or HR director.
Cost: Every hire made against a job description that predates the work arrives needing reskilling from day one. That cost lands in manager time, onboarding duration, and performance review friction, not in a single budget line anyone is tracking.
Pick the role where the gap between the job description and the actual work is widest. Redesign it completely. Name what AI fluency the work requires. Update the hiring criteria. Update the performance measures. Get legal review before it goes live.
Under 20 people: The founder or COO owns this alongside the people lead. The redesign is a two-person conversation: one session to agree on what the role actually requires, one session to write it, one legal review.
20 or more people: The people lead owns this with the hiring manager as the co-author. The hiring manager knows what the work requires today. The people lead knows the criteria, process, and legal constraints. The redesign needs both.
Owner: People lead with hiring manager. Done before the role is posted again.
Cost: Posting the same job description again produces the same hire again. The performance gap does not resolve through iteration on the wrong criteria.
The role redesign is incomplete until the performance measure changes. A job description that names AI fluency as a requirement and a performance review that does not measure it creates ambiguity that compounds every review cycle.
For each redesigned role, add one specific AI fluency measure to the performance criteria. The measure must be evaluable in a review, not a general statement about AI usage. Role-specific and observable is the standard.
Owner: People lead with legal review.
Cost: Without the performance measure, the role redesign is a job description update. With it, it is a process change that compounds. The loop does not close until measurement changes.
The organizations that redesign the role, the criteria, and the performance measure in the same cycle build a compounding loop.
The ones that update the job description and leave the rest unchanged run a campaign that produces the same gap in the next cycle.
What the People Function Owns Now
The gap between what the work requires and what HR measures is a leadership decision. Every quarter it goes unaddressed is a quarter of compounding distance between the organization's capability and the market's baseline.
Engineering teams are building systems that assume AI fluency in the people using them. Operations teams are standardizing on data workflows that require it. Marketing teams are publishing content that depends on it.
The HR process is the last function to catch up. When it does, the gap closes. When it does not, every other function carries the cost of mismatched capability in every hire and every review.
The redesign does not require a new budget. It requires one audit, one role redesigned end to end, and one performance measure updated to match.
The organizations that do all three will hire against the right criteria. The ones that do not will hire against the wrong ones again.
References
1 Microsoft, "London Tech Week 2026: The UK can lead in the AI era," June 2026.
2 TechTimes, "London Tech Week 2026 Closes: Microsoft's $30B UK Push Puts 505,000 NHS Staff on Copilot," June 12, 2026.
