“AI could be the greatest admin assistant in history”

How you can use Artificial Intelligence (AI) in your job, from those in the know.

Lee Murphy, Managing Director, The Accountancy Partnership

AI could be the greatest admin assistant in history. It could do the sorts of tasks needed to check a client’s financial health, or to investigate fraud.

The functionality of AI means it can be applied to many areas including cybersecurity, CRM software, customer service bots – which is where clients and the public interact most with business-powered AI – and more. The benefit is that AI offers a more streamlined customer service experience while, for accountants’ use, it can provide help with financial records.

Incorporating AI

Currently, AI is not part of how we work with our clients’ accounts. We are keeping an eye on the technological and ethical developments with AI and are looking at how we can utilise it in the future – as it will undoubtedly become part of our future. We are a digital practice but we are a community-focused one and we pride ourselves on creating jobs for local people, which is our current focus in these critical economic times.

Risks associated

Integrating AI into an accountancy practice is complex, and staff will need training to use AI to bring a refined nuance to their role. We must be vigilant on data privacy and adhere to a strict ethical code of conduct. The AI itself must be clever enough to work with trained professionals and we must insist on accuracy and a model that offers genuine and useful critical analysis.

There is a theory called technological determination in cultural studies which claims the human race’s decision-making is already largely influenced by technology. The continued development will likely increase our dependence on technology – a topic that needs rigorous scientific and social debate.

Areas AI can help with in accounting

From what we know of AI’s capabilities in accounting, tasks such as trend analysis, reviews of documents, forecasting and planning could all be automated and provided they are closely monitored and the finished result checked, this could be hugely beneficial for accountants to focus on other aspects of business.

James Berridge, Director, Data Analytics, and Becca Durrant, Manager – Cloud Accounting, Saffery

In much the same way as your phone keyboard tries to predict what you’re going to type next, accounting systems with AI could soon be suggesting journals to make based on historic trends combined with current influences.

Better auto-reconciliations should become possible as AI starts to understand the wider context of what’s going on with bank feeds. The recognition of bank transactions has already come a long way since first being released.

Incorporating AI

We are already seeing vendors working on tools to automatically process leases and other legal documents using large language models to interpret and understand the documents. Using machine learning for better predictive analysis could help with improved sensitivity analysis for complex forecasts in audit contexts.

Forecasting and Cash flow apps are a hot topic this year with the need for accountants to take on a more advisory role. There are plenty in the market such as Futrli and Fathom which are using AI to produce predictions based on the company’s historic activities.

Risks

Data privacy and accessibility are key. You don’t want your AI to ‘learn’ from data only certain individuals have access to and effectively leak that information to others. Similarly if your firm’s key data is hidden away in lots of siloed systems then its much harder to benefit from its ability to find and summarise relevant information.

Microsoft has been demonstrating the incredible power of their new M365 Copilot features. It can draft emails and referencing source documentation, which it can pull from other emails, meeting, documents etc – but it requires all your emails, call logs and documents to be stored in a certain way for the AI to access them.

Areas AI can help with in accounting

AI will free up a lot of mundane tasks allowing accountants to focus on the more interesting aspects of their work. Business owners are increasingly looking to their accountants for financial advice in much the same way as they would have previously spoken to their bank managers. AI will leave room for accountants to upskill in advisory roles. The interactivity and data insights provided by AI through many of the apps and the accounting platforms themselves will make these discussions much easier.

Stewart Hurd, VP of Sales, FreeAgent

I don’t believe that we will see a situation anytime soon – if ever – where AI will fundamentally change the way that accountancy is done, or a machine will be able to do the job of an accountant.

The potential of AI really lies in administrative data gathering, management and analysis. This is certainly something that accounting software systems like FreeAgent are building – although it’s unlikely that simple tools like Excel, which have stricter limitations on what they can calculate, will be able to offer them.

In many ways, the future is already here. Methods of automation already envelope us every day and shape the decisions we already make – enabling much of the workflows that you already complete daily, even if in only small ways so far.

We’re going to see a period of acceleration. Machine-learning algorithms are already making banking & expense reconciliation easier, and this kind of functionality will become increasingly sophisticated in the coming years. And AI will become a differentiating factor in leveraging an advantage within that period.

Incorporating AI

FreeAgent has been developing its own AI and automation toolkit since 2018. It has its own:

• bank feed platform, which handled about 300 million transactions in 2023
• AI-driven categorisation which reaches 65% of all transactions from day one, with a 93% accuracy rate
• Intelligent extraction of receipt data to speed up expense creation and automatically match bank transactions
• cash flow forecasting
• benchmarking reports
• customer health scores
• radar insights that collate personalised lists of information and tasks that small business owners need to address in their accounts.

Risks

The risks of AI come when it’s being relied upon for tasks that it wasn’t designed to do. An AI tool like ChatGPT can help you write content based on data that it finds and analyses, but it isn’t sophisticated enough to understand complex issues and put them into the right context. If you’re planning to use that content to inform people (such as your clients or partners), you’ll need someone with expertise to properly review and curate it first.

It’s better to be clear-minded and realistic about the value and scope of AI, rather than be drawn in by the hype.

Areas AI can help with in accounting

• Fraud detection
• risk assessment and management
• algorithmic trading
• sentiment analysis.

Neil Johnson is a freelance business journalist who contributes regularly to trade publications and member organisations, covering employability, recruitment, business trends and industrial analysis.