Our Take
The gap between broadcast and listening is where transformation dies; HR has the position to close it before rollout, not after.
Why it matters
Only 28% of organizations report success maximizing adoption of AI and automation (per APQC). Employees closest to the work see operational realities and process exceptions that project teams miss until after launch, when fixes become expensive workarounds.
Do this week
HR leaders: map workflows with frontline employees and subject matter experts before implementation decisions lock, so you surface adoption risks and process dependencies before pilot begins.
Half of firms broadcast change; few listen first
APQC's global survey of 1,200 HR leaders in Fall 2025 found a striking asymmetry in how organizations approach transformation. About 50.8% regularly communicate the vision of their transformation initiatives (per APQC research). By contrast, fewer than one in five (17.3%) involve employees early enough to shape how change actually lands in daily work.
The result: adoption fails before rollout begins. Only about 28% of respondents report success in maximizing workforce adoption of AI and automation (per APQC), because implementation teams design systems without operational input from the people who will use them. Employees are typically brought into conversations after key decisions about new technologies have already been made.
Employees see what transformation teams cannot
Workers closest to the actual work see process exceptions, customer needs, and workflow dependencies that are invisible to project teams until after launch. Without that insight, organizations discover challenges only after implementation begins. Different teams develop inconsistent approaches to using new technologies. Employees identify customer-facing realities or process dependencies that were overlooked during planning, forcing managers to create workarounds after systems go live.
APQC's research points to a practical solution: involve employees earlier, not to dictate strategy (leaders still set direction), but to contribute operational knowledge before implementation decisions roll out across the organization. Roche, for example, began its Digital Workspace initiative by examining how employees actually searched for information, collaborated, and completed work before designing the solution itself. That understanding shaped a system employees were more likely to adopt.
Early adopters also matter. Novartis created a network of local champions to help employees understand how AI-powered knowledge tools fit into daily work. These champions acted as translators, turning a centralized transformation effort into role-specific conversations. Peer-to-peer learning helps employees move beyond understanding what the technology does to understanding how it fits into their own work.
Measure behavior change, not training completion
Most organizations define successful implementation by training completion rates, communication reach, licenses activated, or rollout timelines. These metrics show whether implementation stayed on track, not whether employees actually work differently.
The more useful question: What behaviors should change if adoption is successful? Depending on the use case, organizations might look for increased usage of new tools, fewer manual workarounds, greater consistency in how teams complete key tasks, or evidence that employees are incorporating new tools into everyday decision-making. One organization in APQC's research measured not just usage rates for its AI-powered search system, but what information employees searched for, where searches failed, and whether users changed their behavior after receiving results. These measures show how employees are interacting with new tools, rather than simply whether they completed training.
The most important adoption decisions occur long before training begins. HR leaders are positioned to influence those decisions by ensuring that employee insight informs implementation plans, adoption strategies, and measures of success before rollout begins.