AI is moving quickly into everyday business. It can draft emails, summarise documents, analyse information, answer questions, generate reports and automate parts of routine workflows. Understandably, this has led to a lot of discussion about productivity and whether businesses will eventually need fewer people.
But focusing only on replacement can cause organisations to miss a much more practical opportunity.
In many businesses, highly capable people spend a surprising amount of their working week doing tasks that do not make full use of their expertise. A finance professional may spend hours gathering information before they can analyse it. A manager may search through emails, spreadsheets and systems to understand the status of a project. A recruiter may repeatedly enter the same candidate information into different platforms. A consultant may spend valuable time formatting reports instead of interpreting what the information actually means.
The problem is not that these activities have no value. They are often necessary. The problem is the amount of human attention they consume.
This is where AI can be useful. Rather than thinking of AI as a substitute for an employee, businesses can think of it as another layer of capability around that employee. AI can help retrieve information, summarise large documents, classify incoming requests, prepare first drafts, identify patterns in data or trigger routine steps in a workflow. The human can then spend more time reviewing, questioning, deciding, communicating and applying professional judgement.
There is an important distinction here between automation and expertise.
A system might be able to identify an unusual transaction, but an experienced finance professional still needs to understand why it happened and whether it matters. AI might summarise a contract, but that does not automatically mean it understands the commercial context or consequences of a decision. It may help a manager prepare a response to an employee, but the manager still needs to understand the person, circumstances and implications.
The greatest value therefore may come from combining the speed of technology with the context and judgement of people.
Businesses should also be careful not to assume that purchasing an AI tool automatically creates productivity. If information is scattered across disconnected systems, processes are inconsistent, access permissions are poorly managed or nobody has defined what the technology is supposed to improve, adding AI can simply create another system for employees to manage.
There is also the question of data. AI can only be useful if it can work with appropriate, reliable information. Before connecting AI to business processes, organisations should understand what information the system can access, where that information goes, who can see the outputs and whether sensitive business or customer information is being handled appropriately.
Human oversight matters for another reason: AI can produce information that sounds convincing while still being incomplete or incorrect. An employee who understands the subject can recognise when something does not make sense. Someone without that expertise may simply accept the output. That makes experienced people more important in many AI-enabled processes, not less.
A useful starting point for businesses is therefore not to ask, “Which jobs can we replace with AI?”
Start by looking for friction.
Where are employees repeatedly copying information between systems? Where are people spending time searching for documents or answers? Which reports require the same manual preparation every month? Which enquiries are repetitive? Where is information available but difficult to access? Which processes create unnecessary administrative work before someone can actually make a decision?
Those areas can reveal where technology may create genuine capacity.
The next step is deciding where human involvement must remain. Tasks involving judgement, accountability, sensitive decisions, complex relationships or significant business consequences should not simply be automated because the technology makes automation possible. A good AI workflow should make it clear where technology assists and where a person reviews, approves or decides.
There is another benefit that receives less attention. When repetitive work is reduced, businesses do not necessarily have to use the saved time simply to reduce costs. That capacity can be redirected.
A salesperson can spend more time with customers. A finance professional can investigate what the numbers actually mean. A recruiter can have better conversations with candidates. A manager can spend more time coaching their team. A specialist can concentrate on difficult problems rather than routine administration.
That is a very different way of measuring the return on AI.
The question is no longer simply “How many hours did we automate?” It becomes “What higher-value work became possible because those hours were returned to our people?”
Businesses considering AI should ultimately be able to answer four questions:
Where are our people losing time?
Identify repetitive administration, information searching, duplicated work and unnecessary manual processes.
Where could technology remove friction?
Look for activities that are structured, repetitive or information-heavy rather than automatically targeting entire jobs.
Where must human judgement remain?
Be clear about decisions requiring expertise, accountability, context, ethics or relationships.
What will people do with the capacity we create?
If saved time has no purpose, automation alone does not necessarily create meaningful business value.
AI will undoubtedly change how organisations operate. Some tasks will disappear, others will change and entirely new ways of working will emerge. But businesses do not have to begin that journey by asking how many people technology can replace.
A far more useful question is:
What could our experts achieve if technology removed more of the work that gets in their way?
That is where AI stops being simply an automation exercise and starts becoming a business strategy.
