I guess that’s probably the challenge that firms are going to face going forward is like, if that basic production work is either done offshore or it’s done using, you know, AI. Then, how do you teach people what good looks like and what the right way of doing things is if they don’t have a chance to do it themselves? Because you’re a trainer. You know this. You learn by doing. You don’t learn by just observing.

John Munden, Chief Strategy Officer at CloudOffis

AI, Automation and the Changing Shape of Accounting Work

There is a lot of conversation about AI in accounting right now. But perhaps the more useful question is not, “Where can I add AI?” It is, “What problem am I actually trying to solve?” I recently sat down with John Munden, Chief Strategy Officer at Cloud Office, ahead of a thought leadership roundtable on the future of the accounting profession. Our conversation explored automation, connected data, changing career pathways, professional judgement and what accounting firms need to think about as technology takes on more of the work we have traditionally done manually.

Don’t bolt AI onto a broken workflow

One of the strongest themes from our conversation was the importance of looking at the workflow before looking at the technology. John’s view is that firms should ideally start with a small group and map how they actually want to work. Where does the work start? Where does it finish? How many touchpoints are involved? Where does the data come from? Who needs access to it? Only then does it make sense to look at the software and automation that can support the process.

I think this is particularly important because accounting firms cannot simply stop operating while they redesign themselves. I have this vision of changing the fuselage of a plane while it is still flying. Clients still need their work completed, deadlines continue to arrive and the team still needs to get through the day. That makes change management just as important as the technology itself. Explain why the change is happening, bring people into the conversation and develop internal champions who understand the software well enough to support others.

Connected data and the single source of truth

Accounting firms have spent years dealing with disconnected systems and data silos. Information sits in one system, gets exported into another and then perhaps lands in a spreadsheet before someone pushes it somewhere else. Every additional step introduces friction and creates another opportunity for information to become outdated.

John talked about the importance of APIs, integrations and particularly two-way syncing. It is not enough to simply pull information into a system. Ideally, changes can also flow back to the original source. Cloud accounting gave us the ability for accountants and clients to look at the same information at the same time, and connected technology should continue moving us towards that single source of truth.

For firms reviewing their technology stack, some useful questions include:

• Can information move easily between the systems we use?
• Does the integration support a two-way sync?
• Can we refresh information without manually rebuilding reports?
• Are we creating another data silo or removing one?

These are not necessarily exciting questions, but they matter enormously if we want AI and automation to work with reliable information.

Vibe coding is exciting, but who maintains it?

We also talked about vibe coding and the increasing ability for people without traditional development skills to create working prototypes. John shared an example of building a business development concept while travelling on a bus, then handing it to his development team. The prototype was not production-ready, but it gave the developers a clear wireframe and potentially saved many hours of development work.

That is a fascinating use case. AI can shorten the distance between an idea in your head and something tangible that another person can see, test and improve. But there is an important distinction between creating something that works today and building something that can reliably operate inside a business for years.

Who maintains the automation when its creator leaves? What happens when an API changes? Where is the data stored? Is the solution secure enough for an accounting firm? John compared this with the spreadsheets and macros that firms have relied on for years. Everything works beautifully until the person who built the spreadsheet leaves and nobody knows how to fix it. AI does not magically remove that problem.

What happens to the traditional accounting career pathway?

Perhaps one of the most interesting parts of our conversation was what automation means for people entering the profession. Traditionally, accountants have learned through a fairly linear progression. You start with basic production work, learn through doing and gradually take on more complex responsibilities.

If AI, automation and offshore teams increasingly handle that entry-level production work, the runway changes. The work may disappear from someone’s daily workload, but the knowledge behind it remains important. How does a junior accountant learn what good looks like if they have never completed the underlying work themselves? How do they identify an error in an automated process if they do not understand what should have happened?

This does not mean we stop developing accountants. It means skills development itself needs to evolve. We may need more scenarios, simulations, structured review and practical learning experiences. Interestingly, AI itself may help us create some of those experiences.

More advisory, but communication matters

For years, we have talked about accountants moving into advisory. The challenge has often been that everyone is simply too busy doing compliance work. If more compliance production becomes automated, that constraint may start to loosen.

But advisory is not simply another software feature. It requires communication. It requires understanding the client, interpreting their information and having meaningful conversations about what happens next. John believes communication skills will become increasingly important as accountants spend more time maintaining relationships and helping clients navigate change.

There is a delightful irony here. AI can write an email and another AI can write the response, but at some point perhaps we should simply pick up the phone and speak to each other. Automating reminders and repetitive communication can be enormously useful if it creates time for genuine human conversations. Automating the relationship itself is a very different proposition.

Professional judgement still matters

Looking three to five years ahead, John expects significantly more compliance production to become automated. Whether every part of that happens within that timeframe remains to be seen, but the direction of travel is clear throughout our conversation.

What remains particularly important is professional judgement. Technology may analyse the numbers, surface information and suggest possible actions. An accountant still needs to understand the context, consider the options and stand behind the advice they provide. In areas such as audit, someone ultimately has responsibility for signing off the work and substantiating where the information came from.

That is why I find this period in accounting so interesting. The opportunity is not simply to do the same work faster. It is to reconsider what work needs to be done by people at all.

If AI and automation can take away repetitive production work, perhaps they give accountants more capacity to interpret, communicate, advise and build relationships. Perhaps the most valuable outcome of all this technology is wonderfully simple: it enables accountants to spend more time actually being accountants.

*Highlighted quote lightly edited for clarity.