Somewhere between answering an email and sitting through another meeting, a rather uncomfortable question has entered the working day. Could artificial intelligence do my job?
It is easy to see why people are asking. AI can draft a customer response, summarise a report and turn a rambling set of notes into something surprisingly presentable. Tasks that once occupied a good part of the morning can sometimes be completed before the coffee has cooled.
Impressive, certainly. Slightly unsettling when those tasks happen to be yours.
AI in the workplace is bringing that conversation closer to home. For businesses, the challenge is deciding which work to hand over, what needs checking and how people fit into a working day that is beginning to look different.
Think about what happens after a customer meeting. Someone needs to organise the notes, record the requirements, prepare a follow-up and make sure the next step does not disappear beneath fifty unread emails.
Few people would describe that as the highlight of their career.
AI can help with much of this preparation. It can turn rough notes into a draft, pull information into a more useful format and suggest a response for someone to review. For a sales representative moving between appointments, that could mean less catching up at the end of the day.
But a tidy summary still needs to reflect what actually happened. Did the customer agree to the deadline, or merely ask whether it was possible? Was that price confirmed? And why has the assistant confidently included a product nobody discussed?
Those details matter when a draft becomes a promise to a customer.
The useful change comes when time saved on preparation gives someone more room to do the work that needs their attention.
This is where the conversation becomes more complicated.
A job usually consists of many different tasks. Some follow predictable patterns. Others involve judgement, responsibility, practical experience or understanding what someone means when they cannot quite explain it.
The International Labour Organization’s 2025 research into generative AI and jobs found that one in four jobs worldwide has some potential exposure to the technology. Its assessment was that transformation is more likely than complete replacement.
That does not mean one in four jobs will disappear. Equally, it is no guarantee that everyone’s role will remain secure or unchanged.
Some tasks will shrink. Roles may be redesigned, and businesses will make different decisions about staffing as the technology develops. Employees will need opportunities to learn, while managers will need to understand what the tools can reliably handle.
Consider customer service. An assistant might answer a straightforward question about an order. A frustrated customer whose delivery has gone wrong for the third time presents a different challenge. Someone still needs to understand the history, decide what can be done and take responsibility for the outcome.
v
Speed is useful. Resolving the problem is what the customer remembers.
For South African businesses, AI in the workplace will also depend on some very familiar technology questions.
Can employees access the tools where they work? Do their phones and laptops support the applications they need? Is connectivity dependable? Do they understand which business information they are allowed to share?
An impressive assistant becomes considerably less impressive when the connection drops halfway through the task.
This brings the discussion into Evercomm’s world of mobile devices, connectivity and managed technology. Choosing technology starts with understanding the person using it, the work they need to complete and the environment they work in.
A sales team travelling between customers has different requirements from employees based at a desk. Introducing AI adds another consideration to those decisions, alongside reliability, security and cost.
Businesses can begin with a manageable task, check the results and ask the people doing the work whether it actually helps. Saving ten minutes is worthwhile. Spending twenty minutes correcting the output deserves another look.
The question of whether AI will take someone’s job will remain with us. For now, there is another question businesses can answer much more directly.
When AI takes something off the working day, what will we do with the time it gives back?
