Published on LinkedIn July 20, 2026
What unsanctioned tools are actually telling you about your AI rollout
This is the fourth issue in a series building on Digital Applied’s “Change Management for AI Adoption: A 2026 Playbook,” published June 14, 2026. Their research framework is the foundation. This newsletter adds the practitioner layer: what these patterns actually look like inside organizations, where the standard response falls short, and what to do first.
78% of employees are using AI tools their employer never approved. That figure comes from a WalkMe and SAP survey of 3,750 enterprise workers, and it is not a cybersecurity statistic. It is a change management one.
Most organizations read it as a governance problem and respond with a policy. That response treats the symptom and misses the signal entirely.
The trust gap nobody is measuring
Prosci measures AI trust on a scale of -2 to +2. Frontline workers land at +0.33. Executives sit at +1.09. That 0.76-point spread between the people authorizing the rollout and the people living inside it is not a communications gap. It is an experience gap.
The WalkMe data makes it concrete. Only 9% of frontline workers say they trust AI for complex, business-critical decisions. Among executives, that number is 61%. The people most exposed to the consequences of poor AI output are those with the least confidence in the tool. The people furthest from daily execution are the most enthusiastic.
This divide does not close with a better announcement. It closes when the people doing the work accumulate enough positive experience with the tool, building trust from the ground up. Most organizations are not designing for that. They are designing for launch day.
Why the leadership support gap is the number that actually predicts outcomes
The trust gap between executives and frontline workers is consequential. The leadership support gap is decisive.
Prosci’s research found that organizations with smooth AI implementations scored leadership support at +1.65 on that same scale. Organizations with struggling implementations scored -1.50. That is a 3.15-point swing on a four-point scale, and it correlates more strongly with rollout outcomes than any technical variable in the data.
The implication is uncomfortable for most sponsor structures. A sponsor who signs the memo, attends the kickoff, and delegates the rest is not providing active leadership support. They are providing nominal cover. The workforce can read the difference between the two faster than any project plan can account for.
Active sponsorship looks different. It means the sponsor uses the tool visibly, talks about what changed for them specifically, and names the initiative in contexts where they could just as easily have stayed quiet. When that behaviour is present, it shows up in the adoption data. When it is absent, the -1.50 score tells you exactly what the workforce concluded.
What shadow AI is actually reporting
The WalkMe survey found that 78% of employees use unsanctioned AI tools, 45% did so in the past 30 days, and 36% did so with confidential data. The standard organizational response is a policy that blocks the tools and a reminder about data governance obligations.
That response answers the wrong question.
When the majority of a workforce is routing around the sanctioned tool, the sanctioned tool is losing a daily individual performance test. People are not bypassing official AI because they enjoy compliance risk. They are bypassing it because the unsanctioned tool does something their official tool does not, faster, more accurately, or with less friction for the specific task in front of them.
WalkMe’s own CEO, Dan Adika, named it directly: “The problem is not AI’s capability. What won’t improve on its own is the trust gap, the governance gap.” The same survey found that 54% of workers bypassed sanctioned AI tools and completed tasks manually in the past 30 days. That number is the more revealing one. Those employees are not choosing a different AI. They are choosing no AI. The official tool lost to a spreadsheet and a keyboard.
Shadow AI is a map of where the sanctioned tool fails the individual test. Mining those patterns, which tasks, which roles, which use cases are driving unsanctioned behaviour, tells you exactly where the change program needs to go next. The organizations treating it as a compliance violation are destroying their most useful diagnostic data.
The practical response to both problems
The trust gap and the shadow AI signal are related. Both are telling you that the individual experience with the tool is not yet good enough to generate the trust and adoption the rollout was designed to produce.
Closing the trust gap requires two things that most rollout plans do not budget for. The first is visible sponsorship that goes beyond the memo, in which leaders demonstrate its use rather than advocate for it. The second is a deliberate effort to generate early wins for frontline employees in their actual workflows, not in a demo environment, because trust is built through accumulated personal experience and cannot be transferred from an executive who already has it.
Reading the shadow AI signal requires treating it as product feedback rather than a policy violation. The questions worth asking are which unsanctioned tools employees are reaching for, what tasks they are completing with them, and why those tasks are not being handled adequately by the sanctioned option. The answers tell you where to focus the next round of enablement, prompt library development, and tool configuration before those gaps quietly drain adoption further.
The organizations that are pulling ahead on AI adoption are not the ones with the most aggressive governance policies. They are the ones that took the trust gap seriously enough to design for it, and treated the shadow AI data as the honest signal it is.
Sources
78% unsanctioned tool usage, 45% in past 30 days, 36% with confidential data, 54% manual workarounds, 9% vs 61% trust for complex decisions: WalkMe / SAP, 2026 Enterprise AI Adoption Survey (3,750 enterprise workers) Building on: Digital Applied, “Change Management for AI Adoption: A 2026 Playbook,” June 14, 2026 https://www.digitalapplied.com/blog/change-management-ai-adoption-2026-overcoming-resistance-playbook
Leadership support +1.65 smooth vs -1.50 struggling, executive trust +1.09 vs frontline +0.33: Prosci, “8 Ways AI-Driven Change is Different” (1,107 professionals) Cited in Digital Applied, June 14, 2026 (above)
Dan Adika quote on trust gap and governance gap: WalkMe CEO, cited in Digital Applied, June 14, 2026 (above)
Strategic Signals publishes monthly on strategy systems and execution clarity for senior leaders. Subscribe to receive each issue directly. Subscribe
If this piece resonated, ALIGN: The Leadership Framework for Clarity in Complex Organizations is available on Amazon. Click Here
