The AI moment in accounting will follow a familiar pattern

In my view, accounting is now at the point mobile and cloud technology once were. There is genuine hype about artificial intelligence, some of it overblown, and there is a great deal of scepticism generated by early tools that make mistakes, pilots that stall, and productivity gains that disappoint on first contact. I don't think any of that is evidence the shift isn't happening. It is what every transition looks like in its unglamorous early years, before the pattern becomes obvious in hindsight.

The more useful question, I think, is not whether artificial intelligence will replace accountants, but which parts of the job change first. Technology tends to automate the simplest, most repetitive tasks before it touches anything else – data entry, categorisation, reconciliation, first-draft reports. That is a meaningful shift for any firm built around producing those outputs, but it is a long way from replacing the work altogether.

Some of the nervousness I hear in the profession is really about value rather than about software. When a client starts to associate a piece of work with the document that comes out the other end, rather than the expertise that sat behind it, the conversation about price gets harder. That is a genuine risk for any firm that describes its services purely as a list of outputs, and it is a reasonable thing for the profession to take seriously.

Picture a business owner who has always assumed she is paying her accountant for a spreadsheet, a return or a report. If a new tool can produce something that looks similar in seconds, she may reasonably start to wonder why the finished document costs what it does. That doesn't mean she stops trusting her accountant. It means the accountant now has to be clearer about what she is actually paying for – the judgement, the context and the accountability that never appeared on the invoice as a line item.

It is not, however, a reason to assume the work itself is disappearing. What does not get automated away is the part of the job that was never really about producing a document in the first place – understanding a client's situation well enough to ask the question they hadn't thought to ask, weighing up a decision before it is made rather than reporting on it afterwards, and being prepared to put a name to advice when the outcome actually matters. Clients have always been paying for that judgement as much as for the paperwork that came with it, even when the invoice never said so directly.

As good intelligence becomes cheaper and easier to access, I expect the market to place less value on producing an answer and more value on knowing whether that answer can be trusted. That shift favours firms and professionals who can show their work is grounded in real context, not just competent output. I see this as a genuine opportunity rather than only a threat: the firms willing to redesign how they work, rather than simply attaching new tools to old processes, are well placed to come out of this transition stronger than they went in.

This kind of change is not new, even if artificial intelligence makes it move faster. Plenty of professions have already been through a version of it, where a digital platform took over search, discovery or the simplest transactions, while the professional kept the relationship, the judgement and the responsibility for the outcome. The ones that adapted well did not try to compete with the platform on the parts it did better. They got clearer, instead, about the parts only a person could still be trusted to do.

One feature of this particular wave of change is worth noting. Earlier transitions needed new hardware, new networks or a slow migration of systems before anyone could really experiment. Artificial intelligence, by contrast, is arriving through software many businesses already have, running on infrastructure that already exists, with an interface almost anyone can pick up immediately. That doesn't remove the real organisational work involved – clean data, sensible controls, retrained staff and rebuilt workflows are all still required – but it does mean the gap between the quiet early years and the obvious later ones may close faster than it has in the past. Firms that wait for certainty before they start preparing may find there is less time than they think.

This is the environment I believe Cloudoffis is building its approach around. For accounting practices, working out where the technology genuinely changes the job, where it does not, and how to adapt without losing what clients actually value them for, is a far more useful conversation than debating whether the profession will survive. It generally will. The more important question, for any individual firm, is what it will look like on the other side, and whether it used the slower early years to prepare rather than waiting for the picture to become obvious to everyone at once.

History suggests the next few years will disappoint the most enthusiastic predictions about artificial intelligence, and the decade after that will exceed them. In my experience, the firms that use the quieter period in between well are usually the ones that come out ahead. That has been true of every technology transition I have watched unfold. There is little reason to expect this one to be different.

John Munden is the chief strategy officer at Cloudoffis.