Healthcare organisations investing in artificial intelligence are running into an older systems problem, with new Strada research suggesting that fragmented workforce technology, spreadsheets, and legacy processes are still shaping the data available to automation and analytics.

Strada says 84% of healthcare respondents continue to use backup systems, spreadsheets, or older processes alongside their main human capital management platform. Two-thirds use multiple systems across HR, payroll, benefits, and finance, while only 27% report a fully cloud-based HCM architecture.

The healthcare-specific findings were released on 18 August and draw on Strada’s broader research into enterprise HCM adoption. The underlying study surveyed 405 senior decision-makers across seven markets, although the public material does not disclose the size of the healthcare subgroup, which limits how confidently the sector percentages can be generalised.

Even with that caveat, the findings point towards a practical constraint on workforce AI. Systems that forecast staffing, identify skills gaps, automate payroll queries, or support workforce planning depend on consistent information across systems that many healthcare employers still operate separately.

Cloud migration does not remove operational complexity

Healthcare employers have spent years moving HR and finance workloads towards cloud platforms, but replacing a core system rarely removes every surrounding application. Scheduling, payroll, agency staffing, benefits, finance, learning, credentialing, and local workforce processes can remain distributed across specialist tools and manual workarounds.

That fragmentation can persist even where the main HCM platform is technically modern. A hospital group may have moved employee records into the cloud while still relying on spreadsheets for rota changes, separate payroll processes for different entities, or manual reconciliations between staffing and finance data.

Strada’s findings suggest that backup processes remain common partly because organisations are reluctant to abandon controls built around operational continuity. Healthcare payroll and staffing are difficult places to tolerate errors, so teams often preserve manual checks even after a new platform is introduced.

The result is a gap between system capability and realised simplification. Cloud software can centralise more data and workflows, although the organisation still has to redesign processes, retire duplicate systems, standardise definitions, and persuade operational teams to trust the new source of record.

Integration determines what AI can see

Workforce AI depends heavily on the quality and consistency of the information it can access. A model asked to predict staffing demand or identify workforce-cost trends can only work reliably if employee status, shifts, absence, skills, payroll, and organisational structures are represented consistently across the underlying systems.

Where those records conflict, automation can reproduce the inconsistency more quickly rather than resolve it. Stephen Dolan, Strada’s senior vice president for healthcare, said AI “can only be as effective as the data behind it”, which is particularly relevant when organisations begin using agents or automated workflows that can act on workforce information rather than simply analyse it.

Integration therefore becomes an operating requirement rather than an IT housekeeping task. Healthcare organisations need to know which system owns a given data field, how changes propagate between applications, which manual overrides remain in use, and whether historical records are comparable enough to support meaningful analysis.

Those questions become harder across large providers with acquisitions, multiple legal entities, union agreements, local staffing rules, or regional payroll requirements. A technically capable AI layer cannot infer organisational rules reliably if those rules are distributed across spreadsheets, human knowledge, and disconnected applications.

Healthcare raises the cost of poor workforce information

Workforce data problems have consequences beyond administrative efficiency because staffing decisions affect service capacity, labour costs, compliance, and patient care. An inaccurate view of available staff or skills can influence rota planning, overtime, agency spend, and whether clinical services have the people required to operate safely.

That makes healthcare a demanding test of the claim that AI can improve workforce management. The sector has large datasets and repetitive administrative processes, but it also has complex roles, regulated working patterns, and little tolerance for automated decisions based on incomplete information.

Strada has a commercial interest in HCM transformation and integration services, so its research should be read in that context. Yet the underlying problem is broader than any one vendor: organisations cannot extract reliable AI value from workforce systems if the information feeding those systems remains fragmented or dependent on manual reconciliation.

Healthcare employers can add AI capabilities quickly compared with the work required to unwind years of system duplication. The more consequential adoption task is likely to be the slower one — making payroll, staffing, skills, and finance data consistent enough that automated systems can be trusted to use it.