Accounting at its core involves the processing of vast quantities of data, which is compiled, collated, presented and used to provide accountability and make good decisions.  

Enter artificial intelligence, or AI. Computer-based automation and cognitive learning are seemingly the perfect match for an industry with a heavy reliance on combining human intelligence with vast quantities of numbers and data. AI in accounting looks to go beyond the likes of process automation and defined calculation to simulating human intelligence as applied in our field.  

Data analysis is a natural first port of call. There is a depth to analysing data with human intelligence. It is one thing for a computer program to execute pre-defined logic over structured data, but another for it to learn what is important through iterating over examples of unstructured information.  

Kira Systems, for example, boasts of machine learning with contract analysis. By viewing existing unstructured documents and contracts, and learning over time the relevance of information to the users reviewing them, it can discern important clauses, identify accuracy and style issues, and highlight them for review.  

Deloitte announced last year that it has been using Kira to efficiently identify and provide insights into client-specific challenges through analysing relevant documents.  

AI and auditing

The audit world is particularly exciting, where fundamental shifts in process such as sampling are possible. Machine learning can, in some cases, be applied to entire databases of data avoiding the need for manual and individual review.  

KPMG maintains that its recently announced partnership with IBM Watson will result in a quantum leap in the type of data used in audit and compliance, and how quickly it is processed.  

"Automated learning could be used to interpret legislative and case law changes and identify where they are relevant to individual clients."

But is it just audit that is a target? Deloitte is already looking at how Kira can be applied in the tax and advisory spheres. There is a huge degree of complexity in interpreting and administering taxation law, but this doesn’t prevent computers from playing a significant role. For example, automated learning could be used to interpret legislative and case law changes and identify where they are relevant to individual clients.  

Advisory is another key application of data analysis. Comparing data across industries and applying it with relevance to clients is already a reality. CA ANZ has been developing CA Kairos, a big data platform available to members that enables the extraction of relevant data and visualisation of the insights.  

Meet your virtual assistant

Moving on from data analysis, another avenue for AI in the accounting sphere is the use of voice recognition technology. MyFirmsApp is planning to use the technology behind Amazon’s Alexa voice-based assistant to develop artificial intelligence apps for accounting. This could take the form of listening to client requests and retrieving information. Voice recognition provides a “face” to the consumer and an ability to interact, while the power of machine learning can provide the content.   

The application of voice recognition technology may prove to be of value in general administrative functions as well. The same technology behind Amazon’s Alexa already powers AI administrative assistants capable of coordinating and scheduling meetings between team members and external parties.  

"By 2021 there will be more texts written by machines than people."

Marry this with AI-based communication, whether it be conveying solutions, drafting letters or just chatting, and you have a powerful personal assistant that can be available 24/7 for your clientele. Robert Dale, chief technology officer at Arria, which produces natural language creation AI technology, suggests that by 2021 there will be more texts written by machines than people.  

Humans in an AI world

The breadth of application begs the question of the role of the human in a world where AI takes on a greater workload. Like most technology, the consensus seems to be that AI will leverage a human’s ability to perform tasks, not entirely replace them.  

“Cognitive enables greater collaboration between humans and systems – so while it’s a game-changer for audit in terms of depth of analysis, it will still require insights from talented people,” KPMG’s National Managing Partner – Audit, Duncan McLennan FCA, told The AFR in 2016. “We’re being helped, not replaced”. 

In the same way automation has reduced the reliance on humans for repetitive tasks, AI has the potential to drastically reduce the reliance on human smarts in a wide range of applications. This should mean that accountants will be freed up to, as Matt Gould from Arria puts it, innovate, model, refine and improve on the expertise of their business.