Current AI usage seems to be a tale of private exploration. A 2023 Deloitte survey found that only 1.4% of Australian businesses and 9.5% of large businesses have officially adopted AI, yet more than a quarter of employees are ‘unofficially’ using it, unbeknown to their managers.

Meanwhile, research suggests 48% of New Zealand organisations use some form of AI in their business, though the rate of official uptake is significantly lower. For example, only 24% say they have legal guidelines in place for use of AI and just 13% have audit assurance.

Having identified efficiency gains, large accounting firms have eagerly bought into automation solutions for tasks such as compliance and auditing.

At the SME level, there is everything from reluctance to pacesetting adoption of AI by firms, says CA ANZ’s senior manager of innovation Jeremy Rowe, who recently launched the AI Starter Toolkit for Accountants webinar. The rate of adoption appears to depend on a leader’s approach to AI.

So, what is a good leader approach? And how are leaders and their firms actually using AI? Two directors share their experience.

NZ firm goes big on automation

In a bid to reduce time spent on manual and repetitive tasks, Wallace Diack Chartered Accountants director Karen Draper CA spent a year testing possible apps. First, she mapped out the client journey from onboarding to payment and asked: where can automation help?

After plenty of testing and trial and error, her firm settled on FYI, a document management and process automation platform. Draper and her team now use more than 150 FYI automations in their day-to-day work.

“When I first started with the firm, I’d spend days doing our invoices and signing documents. Now I don’t do either. Invoices are sent out automatically during the job’s life cycle, and documents are electronically signed using FuseSign, which links to FYI,” Draper says.

It’s a case of minuscule wins in time that accumulate over the week. Saving and uploading documents, creating and sending onboarding documents, engagement letters, emails and work templates makes the little tasks faster. Wallace Diack also developed its own robot in the process: frankie, which automatically downloads client reports and pre-fills workpapers.

All up, Draper estimates these new processes save each staff member about seven hours a week. The firm is also at its highest filing percentage to date with the IR.

“As you get to know each automation, you get faster. I don’t send internal emails anymore. I just push ‘T’ for task and assign it to somebody. And by having a set person assigned to every task, if somebody is away, I can see what they’re working on and assign it to someone else,” says Draper.

With streamlined processes in place, Wallace Diack now pitches for more back-office work with clients. The team is also helping other New Zealand accounting firms to implement FYI in their practices. The current count is more than 20 firms and rising, and it’s easy to see why.

“Instead of having someone spend 18 months building these automations, we can do it for them in two days. It’s how you build up and automate that is key,” Draper says.

“Having a clear understanding of your business processes is critical. Once you do, you can wrap an automation around it and reduce time spent on manual work, increasing the quality of the work and your firm’s capacity.”

A ChatGPT pacesetter

James Carey CA is a director of North Sydney firm, Prime Partners. Carey has been dabbling with ChatGPT with a trial-and-error approach that has yielded some neat results.

His big coup was an internal reporting dashboard for his 30-plus staff. Previously, the team’s internal reporting process was ad hoc and manually based, requiring administrative power and hours. Carey sought ChatGPT’s help to create a lookup function that would enable monthly reports for an individual to be called up on the spot.

“Given time, I could probably get there myself by googling it or looking around in forums, but it would be a frustrating experience. So, I fed the structure of the data into ChatGPT and told it the outcome I wanted and asked: how do I achieve that?” Carey explains.

“There was some back and forth – the answer is rarely right the first time – and eventually it gave me a very long Excel formula. The formula may not be elegant, but it works. The dashboard has been a big benefit for the team.”

In addition to handy Excel formulas, Carey uses ChatGPT to manipulate data in concert with Python, a programming language he became aware of through a CA ANZ course. He asks ChatGPT for Python scripts to perform menial and otherwise lengthy tasks, like encrypting hundreds of PDFs received from the ATO, cleaning up general ledger data provided by a client in CSV form, and compiling scores of Xero invoices into one PDF and ensuring there is one invoice per page.

“The invoice solution was a 15-minute exercise of going back and forth, but it would have been a two-hour exercise for someone to go through and split it up manually. It saves money for the client because they’re not paying for us to manually process data,” says Carey.

Data safety is an understandable sticking point for CAs when it comes to using generative AI. To avoid any security issues, Carey tells ChatGPT the structure of the data or gives it dummy data – he never provides the tool with client data. Sharing of data can be turned off in ChatGPT’s settings, but Carey’s risk-free approach will undoubtedly appeal to other CAs.

---

ChatGPT uses for accountants

● Research a client or industry

● Come up with Excel formulas

● Software training (e.g. Excel, Python)

● Data manipulation

● Help draft client communication (emails, letters)

● Help create marketing content for social channels, blogs and newsletters.

---

Top accounting AI technology

Aider.ai: prepare for client meetings, generate sales scripts and analyse client data

Etani.ai: report on staff and billable hours

XBert.io: automatically check invoices and GST.

---