Across industries, businesses are putting approved AI tools into the hands of employees, but many employees are already choosing their own.
Microsoft research found that 71% of UK employees have used unapproved consumer AI tools at work, and more than half use them every week.
For 41%, these are tools they already use in their personal lives, while more than a quarter say their employer offers no approved alternative.
This is the Shadow AI economy. It creates an obvious security challenge, but it also shows where demand already exists.
Employees are choosing these tools because they are finding practical value in them, often before their employers have caught up.
Why Shadow AI persists
Providing an approved AI tool doesn't automatically change the habits employees have already formed. Consumer AI has set expectations around speed, flexibility and ease of experimentation. Employees can try different approaches, move between tasks and use the technology for whatever they need to get done.
Approved enterprise AI has to meet a different set of requirements. It needs to protect company data, respect access controls, integrate with existing systems and operate within governance requirements. But employees still expect it to be just as useful and easy to use.
If the sanctioned option is more restrictive, harder to use or less suited to the way work gets done, employees are likely to return to the tools that work better for them. Blocking those tools may reduce the immediate risk, but it doesn't make the approved alternative any more useful or address why employees chose unapproved tools in the first place.
Shadow AI can therefore show businesses where their approved tools are falling short of what employees actually need.
Shadow AI is a signal of where demand already exists
We can't overlook the risks that come with Shadow AI. When employees use personal accounts or unapproved tools, organizations can lose visibility into what company information is being shared, how it is being processed and where it may go. But if employees are choosing these tools despite those restrictions, businesses should pay attention to what they are using them for.
If employees repeatedly turn to AI to analyze information, create content or speed up work that would otherwise take them hours, they are already showing where AI is useful in their day-to-day work.
We saw this in our own organization. Employees were already using tools such as ChatGPT and Claude to create presentations, draft content and produce other everyday assets. We looked at where they were saving time and brought those tools into an approved enterprise environment.
We also created shared skills that employees across the organization could use. Presentations are a good example. A shared skill in ChatGPT or Claude can include our enterprise messaging, brand language and the context employees need to talk about the company accurately. Anybody across the organization can create a presentation deck with standard brand guidelines and style.
For us, this has been the balance to get right. Employees can continue using tools they find useful, with company context and shared skills built in. At the same time, we have visibility into how those tools are being used and how company data is flowing through them, with enterprise security and access controls around that activity.
Businesses can take a similar cue from their own employees. Rather than trying to predict every AI use case centrally, look at where employees are already turning to AI and bring what is working into an approved environment.
Build beyond today’s AI tools
Employees may favor tools like ChatGPT, Claude or Gemini today, but that will change. New tools will come along, and existing ones will get better. Businesses cannot start again every time employees find something new that works for them.
The controls need to outlast the tools. Access to company data, what employees can do with it and the visibility the business needs should not depend on which model someone uses. Put those controls across the AI environment, and the business can introduce or replace tools as employee needs change without having to rebuild everything around them.
Also, businesses do not need to predict which tool employees will choose next. They need to get better at recognizing what is useful, bringing it into the business and making it safe to use. That way, the next tool employees discover does not have to become the next Shadow AI problem.
Protect your business with the best antivirus software.
This article was produced as part of TechRadar Pro Perspectives, our channel to feature the best and brightest minds in the technology industry today.
The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/pro/perspectives-how-to-submit

