Shadow AI is often framed as an employee problem. But industry research finds high rates among senior leaders, and Larridin Spark data shows why executive AI usage deserves a closer look.
The Short Answer
Shadow AI isn't limited to frontline employees. TrustedTech found that 65% of senior decision-makers used unapproved AI tools, compared with 31% of other employees. In one Larridin customer, 76% of the Leadership/Executive group's tracked AI interactions involved tools classified as unapproved, up 302% from the prior period. Some of that activity may reflect desktop-client classification issues, so the next step is to investigate the tools behind the number.
Key Wins
What’s a Spark Insight?
A Spark Insight is a data-backed finding Larridin surfaces automatically from how your teams actually use AI, flagging the patterns worth a leadership conversation.
Who This Matters To
CIO · CISO · CAIO
| If you’re asking… | The answer | Go deeper |
|---|---|---|
| Is shadow AI common among senior leaders? | Yes. TrustedTech reported unapproved AI use among 65% of senior decision-makers versus 31% of employees below decision-maker level. | The research |
| Are executives aware of shadow AI in their company? | Executive confidence doesn't necessarily reflect actual visibility. One study found 90% of executives confident in their AI visibility while 52% of employees used unapproved tools. | The research |
| How high can executive shadow AI rates get? | In one Larridin customer, 76% of the Leadership/Executive group's AI interactions involved tools flagged as unapproved, up 302%. | The data |
| Is every “unapproved” AI tool really shadow AI? | No. Desktop versions of approved tools are sometimes classified as unapproved, so high rates are a reason to look closer before acting. | Read the numbers carefully |
AI governance programs tend to look down the org chart: train employees, block unapproved tools, monitor usage. But leaders face pressure to move fast with AI.
| What you’re struggling with | What you actually need | How Larridin helps | Evidence |
|---|---|---|---|
| “Our AI policy is mostly aimed at employees.” | Visibility into AI use at every level | Measures unapproved AI usage by department, including leadership | One org: leadership unapproved rate 76% |
| “Executives think we have AI under control.” | Data to compare confidence with reality | Shows actual unapproved usage alongside approved tools | Okta: 90% of execs confident vs. 52% of employees using unapproved tools |
| “Leadership shadow AI is growing fast.” | Trend data by group | Tracks growth in unapproved usage over time | One org: leadership unapproved usage up 302% |
| “I’m not sure the numbers are accurate.” | Tool-level detail to separate real shadow AI from misclassification | Breaks unapproved usage down by tool and client type | Desktop clients of approved tools often show as unapproved |
The question you’re asked: “Where is our shadow AI risk concentrated?”
What to do: Break unapproved usage down by department, including leadership. Expect the top of the org to need governance as much as anyone.
For more on enterprise-wide AI visibility, see AI Monitoring for CIOs.
The question you’re asked: “How do I raise shadow AI with executives?”
What to do: Lead with data and fixes, not blame. Offer approved alternatives that meet leaders’ need for speed.
For more on governance visibility, see AI Monitoring for CISOs.
The question you’re asked: “Why are leaders going around our approved tools?”
What to do: Ask what the approved tools don’t do for them. Unapproved AI use may point to a gap in the approved stack.
For more on managing AI adoption and oversight, see AI Monitoring for CAIOs.
Because senior leaders report high rates of unapproved AI use, even though they're responsible for organizational oversight.
TrustedTech's research found that 65% of senior decision-makers used unapproved AI tools, compared with 31% of other employees. Among C-suite respondents specifically, the figure reached 73%.
Larridin Spark data points the same way. In one customer, the leadership and executive group’s unapproved AI usage rate reached 76%, up 302%.
Myth: Shadow AI is mostly a frontline employee problem.
Reality: Research shows higher reported use among senior leaders, and Larridin data identifies executive-level activity worth investigating.
The survey doesn't explain why, but it reveals a gap between executive confidence and employee behavior.
Okta's 2026 research found that 90% of executives were confident in their organization's AI-tool visibility, yet 52% of workers reported using unapproved AI tools. The findings compare two survey groups, not measured usage at the same companies. They're a reason to validate executive confidence with behavioral data.
As a reason to look closer, not a final verdict.
Larridin data shows that desktop clients of approved tools are often classified as unapproved. Before treating a high rate as a policy violation, check which tools drive it. Some will be real shadow AI. Others will be approved tools showing up under a different client.
Shadow AI is the use of AI tools that haven’t been approved by the organization, often through personal accounts or browser extensions.
Leaders handle sensitive information, such as strategy, financials, and personnel matters, so unapproved tools at that level can expose more.
Make approved tools meet their needs, give them visibility into their own usage, and apply the same policy at every level.
See unapproved AI usage by group, including leadership, and bring it to your next security review.