The numbers we use, and where they come from.

Fewer than a third of HR leaders say their workforce data is reliable enough to build AI models on.

Gartner, HR Technology Survey 2026External source

The average enterprise runs HR data across five or more systems that do not share a common employee identity.

Sapient Insights, HR Systems Survey 2025-2026External source

HR teams spend roughly a third of reporting time reconciling headcount numbers between systems rather than analysing them.

Josh Bersin Company, 2026External source

About four in ten HR functions have at least one AI agent in production, up from one in ten two years earlier.

McKinsey, State of AI in the Workplace 2026External source

Employment-related AI use cases are classified high risk under the EU AI Act, requiring logging and human oversight.

European Commission, AI Act implementation guidance 2025External source

Fewer than half of HR AI pilots reach company-wide rollout within a year of launch.

Deloitte, Human Capital Trends 2026External source

Voluntary attrition models built on tenure, manager change and compensation data typically explain 60-70% of resignations.

Visier, Workforce Benchmark Report 2025External source

Replacing a mid-level employee costs on average six to nine months of their salary in recruiting and ramp-up time.

SHRM, Cost of Turnover Study 2025External source

Contractor spend and overtime commonly add 15-20% on top of planned workforce cost by year end.

Mercer, Workforce Planning Survey 2026External source

Only around a quarter of finance and HR leaders review the workforce plan and the financial plan on the same numbers each quarter.

PwC, Workforce of the Future Survey 2026External source

Roughly 40% of core job skills are expected to change within five years across most industries.

World Economic Forum, Future of Jobs Report 2025External source

Organisations with a live skills taxonomy fill internal roles about twice as fast as those without one.

LinkedIn Talent Solutions, Workplace Learning Report 2026External source

A tool with no measured adoption rate has a median usage of under 20% six months after rollout.

Gartner, Digital Worker Experience Survey 2026External source

Manager enablement is the strongest predictor of whether an HR tool survives past the pilot stage.

Josh Bersin Company, 2025External source

Most HR functions evaluating AI vendors cite data readiness, not the model itself, as the main blocker to going live.

Deloitte, Human Capital Trends 2026External source