What’s driving UK employees to rely on unapproved AI tools, and what does this say about gaps in workplace systems today? We spoke to Alexey Korotich, Chief Product Officer at Wrike for his exclusive insights on how employers can tackle the issue of shadow AI in the workplace.
Here’s what he told us:
Employees are turning to external tools to keep up with the pace of work. Our latest research found that 42% of UK workers have turned to unapproved AI tools, most often to complete tasks faster than internal systems allow. This signals that many organisations still rely on slow, fragmented systems that don’t support how teams actually operate. When approved internal tools create friction, employees naturally seek out alternatives that are faster and more intuitive, even if it means bypassing policy.
This behaviour exposes a gap not just in workplace technology stacks, but in workplace strategy. Employees are eager to embrace automation, but without clear guidance or connected data, they risk reinforcing silos and introducing security risks. If organisations want to reduce shadow AI, they must meet employees where they are at by providing responsive, approved solutions that support real-time execution and protect institutional knowledge.
How significant is the security and compliance risk posed by shadow AI, and have you seen real examples of these risks materialising?
The risks tied to shadow AI are very real, primarily because leaders lose visibility into how workplace data is being handled. Once sensitive information is entered into an unapproved AI tool, it can be stored, processed or reused in ways the organisation can’t monitor or control. While leaders lose visibility into how data is handled, that data security issue is merely a symptom of a much deeper, structural problem. When employees use unsanctioned tools, workflows become fragmented and valuable institutional knowledge gets scattered. This creates confusion, duplicates effort, and ultimately undermines the very efficiency employees were trying to achieve.
There are cases where teams are under pressure to move quickly, so they turn to unsanctioned tools and unintentionally expose sensitive information. These acts are often the result of system gaps. When documents are scattered, and guidance is unclear, employees default to the fastest solutions rather than the most secure ones.
What steps should organisations take in 2026 to build AI governance frameworks that actually match how employees work day-to-day?
In 2026, successful organisations will embed secure frameworks and AI capabilities directly into the systems employees already use to manage work, set priorities, and collaborate. By building guardrails directly into employee platforms, instead of relying on separate policies or disconnected oversight, leaders can deliver intuitive and consistent governance. This also requires transparency, so employees understand how decisions are being made and feel confident in the outcomes. By establishing strong governance frameworks that are integrated into daily workflows, leaders in 2026 will help remove uncertainty around AI and become a driver of trust in the technology and how an organisation is managing it.
How can leaders encourage safe, approved AI adoption without slowing down innovation or creating unnecessary barriers for employees?
Leaders must shift their mindset from restriction to enablement. The debate isn’t about safety versus speed; it’s about creating an environment where the safest path is also the fastest. People will prioritize speed over process. The goal is to make the approved path that path.
This requires three core actions:
-
Outcompete shadow IT. Make your approved tools undeniably better. They must be powerful, intuitive, and seamlessly embedded into existing workflows so that turning to an outside tool feels like a downgrade, not a solution.
-
Educate, don’t just mandate. True governance comes from understanding, not just rules. When employees grasp the ‘why’ behind the security boundaries, and see the value in the approved tools, responsible use becomes second nature.
-
Govern at the speed of innovation. AI evolves in weeks, not years. An annual governance review is already obsolete. Leaders must create an agile framework that adapts to new technology as it emerges. If your internal systems can’t keep pace, shadow AI will always fill the gap.
Ultimately, you don’t slow down innovation to maintain safety. You channel innovation through superior infrastructure, making the “right way” the only way your team wants to work.
What does Wrike’s data reveal about the types of AI tools employees are turning to most, and what does that indicate about future AI needs in the workplace?
Wrike’s data shows that employees are gravitating toward AI that helps them move faster with tools that support task execution, prioritisation, and automation. In the U.K., 86% of knowledge workers report using AI tools at work, most commonly generative AI (66%), copilots or assistants (56%), and intelligent agents (36%). These tools are increasingly woven into how teams plan, execute and manage work.
But adding more tools hasn’t guaranteed more value. Employees are now looking for connected solutions that help encourage faster execution and easier access to accurate information. Nearly all U.K. respondents (99%) say there is value in tools that can share context and coordinate in the background.
This points to a clear shift where the future of workplace AI lies in integration. Employees want what we at Wrike call ‘connected intelligence’ or intelligent support woven directly into the systems where work already lives, with contextual awareness of projects, systems, and priorities. To meet this need, organisations must move beyond fragmented AI rollouts and invest in unified platforms that embed intelligence directly into workflows.