A board question your dashboards can't answer yet
You have the deck built. Headcount by department, voluntary turnover trend, engagement scores, time-to-fill by function — the quarterly workforce analytics package your board has seen a dozen times. Then someone on the audit committee, or the PE partner who sits in twice a year, asks the question that wasn't on last quarter's agenda: which of our roles are most exposed to AI, and what's the plan?
Your workforce analytics software — whatever combination of HRIS reporting, BI dashboards, or a dedicated people-analytics platform you've assembled — has no answer. It was built to describe the workforce you have, not to score how exposed that workforce's task mix is to a technology curve that's moving faster than your last platform refresh.
This guide covers what workforce analytics software is actually built to measure, where the AI-exposure question sits relative to that, and how a mid-market HR team can combine both layers into a single planning cycle without buying (or building) something oversized for the problem.
What workforce analytics software actually measures
Most platforms in this category — from HRIS-native reporting to dedicated tools — are organized around a common set of questions: who do we have, where are they, what do they cost, and are we keeping them. That typically means:
- Headcount and org structure — current state, historical trend, span of control
- Turnover and retention — voluntary/involuntary splits, tenure cohorts, flight-risk modeling
- Compensation and cost — pay equity, comp ratio, cost-per-FTE by department
- Engagement and sentiment — survey results, pulse data, sometimes attrition prediction
- Time-to-fill and pipeline health — recruiting funnel metrics
This is descriptive and, in the more advanced platforms, predictive analytics on outcomes the organization has already experienced — who tends to leave, who tends to get promoted, where cost concentrates. It answers "what is happening" and, in the better tools, "what is likely to happen next based on our own history." For a deeper walkthrough of this category and how to evaluate it for a mid-market HR team, see our guide to HR analytics tools.
What none of it answers, by design, is a forward-looking structural question: which of the tasks our people actually do are, today, well within reach of generative or agentic AI tools — regardless of whether anyone has left, been hired, or expressed an opinion in an engagement survey.
The blind spot: exposure to AI, not just historical performance
The research on this point is now fairly dense, and it's worth citing precisely because it is not company-specific — which is exactly the limitation a mid-market HR team needs to understand before treating any of it as a plan.
McKinsey's 2023 analysis estimated that 60–70% of employees' time could theoretically be automated with current generative AI capabilities, up from roughly 50% in prior estimates. Goldman Sachs, the same year, put a jobs-equivalent number on the same dynamic: 300 million full-time jobs, globally, exposed in some degree to generative AI substitution or complementarity. The World Economic Forum's Future of Jobs Report 2025 frames the churn more structurally: 170 million jobs created and 92 million displaced by 2030 — a net addition of 78 million, but 22% total churn, with 39% of core skills expected to transform or become outdated in the same window.
These are macro, sector-level, or economy-wide figures. None of them tell a VP of People at a 900-person logistics company which of their roles carries that exposure, because none of them were built to. That's the layer traditional workforce analytics software also doesn't touch — and it's a genuine gap, not a marketing framing. Gartner's 2024 research on strategic workforce planning found that 86% of HR leaders have not implemented strategic workforce planning, and a separate Gartner survey the same year found 66% say their workforce planning is limited to headcount planning and struggle to demonstrate its ROI.
Gartner, 2024: 86% of HR leaders have not implemented strategic workforce planning — the gap between "we have workforce data" and "we have a workforce plan that accounts for AI exposure" is where most mid-market teams currently sit.
That gap is exactly where an AI exposure layer, applied at the task level and rolled up by role and department, is designed to sit — not as a replacement for headcount and turnover analytics, but as an additional input those tools were never built to produce.
Inside the four-dimension exposure rubric: a worked example
To be useful to a board, an exposure score has to be explainable to someone who will ask "how did you get this number." The methodology WorkforceAnalysis uses starts from the U.S. Department of Labor's O*NET database — a public taxonomy covering 1,016 occupational titles (923 at the data level, mapped to 55,000+ real-world job titles), maintained with roughly 277 descriptors per occupation and updated on a regular cycle, primarily each Q3. Each occupation in O*NET is broken into detailed work tasks, each carrying an importance weighting — how central that task is to the occupation as a whole.
This site incorporates information from ONET. Used under the CC BY 4.0 license. ONET is a trademark of USDOL/ETA.
Every task is then scored across four dimensions, each on a 0–100 scale:
- Cognitive routine — how standardized and rules-based the reasoning involved is
- Physical routine — how repetitive and specifiable the physical actions are
- Social/judgment — how much the task depends on human relationship, negotiation, or contextual judgment
- Creative — how much the task requires original, non-formulaic output
Here's a simplified worked example, using round numbers for illustration. Take a role built from three O*NET tasks, weighted by importance:
| Task | Importance weight | Cognitive routine | Physical routine | Social/judgment | Creative |
|---|---|---|---|---|---|
| Reconcile monthly reports | 0.5 | 80 | 10 | 15 | 10 |
| Present findings to stakeholders | 0.3 | 30 | 5 | 75 | 40 |
| Design a new tracking template | 0.2 | 40 | 5 | 20 | 70 |
Weighting each dimension by task importance and summing gives a role-level exposure profile — in this illustration, a cognitive-routine score around 57, a social/judgment score around 30, and so on. That role-level composite, not a single "AI risk" number, is what feeds a department heatmap and a category assignment: Monitor, Review, or Redeployment Candidate.
None of those three labels is a verdict. A Redeployment Candidate designation means the task composite warrants a structured look at internal-mobility options grounded in the same O*NET occupational network — not that the role, or the person in it, will be eliminated. Exposure scoring is an input to a human planning conversation, not a prediction of who keeps their job. That distinction is the entire point of building the rubric transparently rather than delivering a black-box score.
Where platforms like Visier stop and exposure analysis starts
Visier is a reasonable reference point here because it sits at the more analytically sophisticated end of the workforce analytics category: a people-analytics platform built for mid-to-large enterprises, with predictive analytics and a library of pre-built workforce metrics, and — like most platforms at this end of the market — it does not publish standard list pricing.
What Visier's platform is not built to do is task-level AI exposure scoring against the O*NET occupational taxonomy, and it does not include a redeployment-recommendation engine tied to that scoring. That's not a criticism of the product; it's a scope difference. Visier and similar enterprise platforms were designed to answer "what's happening across our workforce," using the organization's own historical data as the predictive base. Exposure analysis answers a different, narrower question — "how exposed is the actual task composition of this role, today, based on an external and stable public taxonomy" — and was never something the broader people-analytics category set out to cover.
For a mid-market HR team, the practical implication is that you likely don't need to replace your existing workforce analytics software to get an exposure layer. You need to add one. If you're still comparing platforms at the category level, our guide to people analytics software for the mid-market covers how to evaluate fit before you add anything new to the stack.
Building one planning cycle from both layers
The two layers work best read together, on the same cadence:
- Descriptive/historical layer (your existing workforce analytics software): current headcount, turnover trend, cost-per-FTE, engagement — tells you the state and trajectory of the workforce you have.
- Exposure layer: task-level scoring by role, rolled up by department, against a public, explainable taxonomy — tells you where the structural task composition of the work itself is concentrated toward the kind of routine cognitive or physical work current AI tools handle well.
Read alone, historical analytics can tell you a department is stable and well-engaged while its task composition is heavily weighted toward the cognitive-routine end of the rubric — a combination that looks fine on this quarter's dashboard and warrants a very different conversation on a three-year horizon. Read alone, exposure scoring says nothing about your actual attrition risk, cost structure, or engagement — it was never built to. Together, they let a People Analytics Lead walk into a board meeting with both what's true today and what the underlying task structure suggests is worth monitoring, without conflating the two.
This is also where strategic workforce planning stops being a headcount exercise, which is the exact gap Gartner's research flagged above. Our guide to strategic workforce planning with AI walks through building that combined planning cycle in more detail.
Choosing and piloting workforce analytics software for the mid-market
A few practical filters, if you're evaluating either category right now:
- Match the tool to the question, not the category. If the immediate ask from your board or ownership is specifically about AI exposure, an enterprise people-analytics platform is the wrong first purchase — it wasn't built to answer that question, regardless of its sophistication elsewhere.
- Check whether pricing scales to your headcount. Mid-market teams (200–5,000 employees) are frequently quoted enterprise-platform pricing structured for organizations several multiples larger. Ask directly whether the tool has a self-serve or mid-market tier before a sales cycle.
- Ask for the methodology, not just the output. Any exposure score you can't explain to a CFO is a liability in a board setting, not an asset. Insist on seeing the underlying rubric and data source, not just a dashboard.
- Plan for both layers, not one. Budget and roadmap for descriptive workforce analytics and forward-looking exposure analysis as complementary line items, not competing ones.
If you want a structured starting point for that second layer, our pricing page outlines the tiers WorkforceAnalysis offers for organizations at this size, and the store includes the HR AI Strategy Toolkit — a downloadable template built to help HR strategy teams structure the board conversation once the exposure data exists, covering how to frame Monitor, Review, and Redeployment Candidate findings for an executive audience without overstating what the data shows.