AI Productivity in the Workplace: Are Employees Going At It Alone?
Employees are already embracing AI at work. But what happens when they aren't given the tools, training or guidance they need?
Artificial intelligence has moved rapidly from something organisations were experimenting with to something employees are using as part of their everyday working lives.
But a surprising development is emerging: employees aren't always waiting for their organisation to provide AI. Some are paying for it themselves.
New research from Deloitte UK, based on responses from 25,000 UK workers, found that 17% of people who use generative AI for work pay for at least one AI tool themselves. Deloitte estimates that UK workers are spending around £958 million a year on GenAI tools for work.
For organisations, this isn't simply an interesting statistic about AI adoption, as it actually raises a much bigger question:
Are employees finding their own ways to become more productive because organisations haven't yet given them a clear, supported way to use AI?
AI Productivity in the Workplace Is Already Happening
The demand for AI productivity in the workplace is clearly coming from employees themselves.
Deloitte's research found that 63% of UK working adults aged 18–70 knowingly use GenAI for work, while nearly a quarter use GenAI tools every day for work. Among GenAI users, the most common uses include searching for information, drafting emails and creating summaries.
These are exactly the kinds of everyday activities where employees are looking for time savings. They want to:
Find information faster
Draft emails more quickly
Summarise information
Create content
Reduce repetitive work
Get through their workload more efficiently
And many are prepared to find their own solutions.That's a positive signal about employees' willingness to embrace new technology. But it also creates a challenge for organisations.
Individual experimentation can move faster than organisational policy, training and governance.
The Rise of Shadow AI
When employees use AI tools without their organisation's knowledge or approval, this is often referred to as Shadow AI.
Deloitte found that 31% of GenAI users in the UK say they use GenAI without their employer's knowledge. Free tools are also widespread, with 46% of GenAI users reporting that they use free tools at work.
This creates an uncomfortable situation for organisations.
Employees may be using AI because they genuinely believe it helps them work better. Yet the organisation may not know:
Which tools are being used
What information is being entered
How AI-generated content is being checked
Whether confidential information is being shared
How much time is actually being saved
Whether different teams are developing completely different ways of working
The answer isn't necessarily to tell employees not to use AI. The bigger opportunity is to understand why they are turning to it in the first place.
Why Are Employees Paying for Their Own AI?
Deloitte's research provides some useful clues. Among workers who pay for their own GenAI tools or use tools that may not be approved, 21% say the tools outperform company-provided tools, while 14% say the tools are essential to their job but aren't funded by their employer.
This suggests that some employees aren't simply chasing the latest technology, they are trying to solve real problems.
They may have discovered that an AI tool helps them write faster, find information more easily or manage part of their workload more effectively.
For organisations, that should prompt a different question from simply asking:
“How do we stop employees using unapproved AI?”
Instead, organisations should be asking:
“What are employees trying to achieve with AI, and how can we help them achieve it safely and productively?”
That shift in thinking could be important for AI productivity in the workplace.
Providing AI Tools Is Only the Beginning
There is an understandable temptation to believe that the answer is simply to provide employees with an approved AI tool, but access doesn't automatically create productivity.
Deloitte found that around half of employees who use GenAI at work have received no formal training on how to use AI safely and effectively. It also found that many users are applying AI to familiar activities such as searching and writing emails rather than using it for more transformational work.
This is where organisations need to think beyond technology.
Giving someone Microsoft 365 Copilot, for example, doesn't automatically teach them:
What they should prioritise
How they should manage their workload
Which tasks should be automated
How commitments should be tracked
How information should be organised
When AI should be used
When human judgement is essential
How their team should work consistently
AI can speed up a process. However, it doesn't necessarily create a good process.
AI Productivity in the Workplace Needs Better Ways of Working
This is an important distinction.Imagine someone has an overloaded inbox, a calendar full of meetings and dozens of competing tasks.
Copilot can help summarise emails. It can draft responses, identify actions, prepare for meetings, the list goes on. It doesn't however, decide which commitment should come first.
It doesn't establish how the team manages priorities, create clear ownership of work or decide whether another meeting is actually necessary.
Those are ways-of-working decisions, in other words, humans decisions.
This is why AI productivity in the workplace isn't simply about implementing AI. It is about combining AI with practical, consistent working practices.
Microsoft 365, Copilot and the Missing Link
Microsoft 365 already provides organisations with a powerful collection of tools for communication, collaboration, planning and information management.
Copilot adds another layer of capability by helping people work with the information and content within those tools.
But there is still a missing piece:
How do people actually work?
WorkingSm@rt® addresses this by focusing on the processes and behaviours that sit behind the technology. The aim isn't simply to teach people more Microsoft 365 features.
It's to help people establish better ways of working around:
Priorities
Time
Tasks
Email
Meetings
Delegation
Collaboration
Information
Projects
Follow-through
When those foundations are in place, AI has something much more useful to accelerate.
Microsoft 365 provides the tools. WorkingSm@rt provides the ways of working. Copilot provides the acceleration.
From Individual AI Experiments to Organisational Productivity
The fact that employees are spending their own money on AI should not only be seen as a technology or security issue. It is also a signal.
Employees are telling organisations that they want to find better ways to work.
The challenge is turning that individual enthusiasm into organisational productivity.
That means creating clear expectations around AI use, providing appropriate tools, giving people practical training and establishing consistent ways of working.
Deloitte's research reinforces this point: its findings suggest the challenge is no longer simply getting people to use GenAI, but helping them use it in a way that is secure, responsible and creates lasting value.
What Should Organisations Do Next?
Organisations don't necessarily need to introduce another AI initiative.
They should start by understanding how work is currently being done.
1. Find out how employees are already using AI
Ask people what tools they use and what problems they are trying to solve.
The answers may reveal opportunities that aren't visible from the top down.
2. Provide clear guidance
People need to understand which AI tools they can use, what information can be shared and where human judgement remains essential.
3. Train people to use AI effectively
Training shouldn't only focus on prompts and features.
It should connect AI to the real work employees need to accomplish.
4. Establish better ways of working
AI works best when people already have clarity around priorities, tasks, commitments, meetings and information.
5. Measure outcomes, not AI activity
The important question isn't how many prompts someone has written.
It's whether they are saving time, focusing on higher-value work, improving collaboration and delivering better results.
The Future of AI Productivity in the Workplace
Employees are already voting with their time, and sometimes their wallets.
They are experimenting with AI because they believe it can help them work better.
Organisations now have an opportunity to catch up.
The goal shouldn't be to stop employees experimenting with AI. Nor should it simply be to give everyone access to another tool.
The opportunity is to create a clear, supported and consistent way of working with AI.
Because the real question isn't:
“Are our employees using AI?”
It's:
“Are our employees working better because of AI?”
And that's where better ways of working become the missing link between AI adoption and real productivity.
Priority Management helps organisations build those better ways of working across Microsoft 365, including Copilot. Lear more about our training course below.
About us > Priority Management is a worldwide training company with 55 offices in 15 countries. We have successfully trained more than two million graduates in Priority workshops. Our programs help companies and people be more effective and manage their workflow in and out of the office by providing tools, processes and discipline.
Simply put - A Better Way To Work! Clients range from Fortune 500 companies, small-to-medium businesses and government/military employees. Click Here to view all our courses.
Found some useful insights? Share them with someone you care about ↷
Comments