Jamie Renehan FCCA leads Bank of Ireland’s advanced analytics team

As the digital transformation of accounting and finance gathers momentum, artificial intelligence (AI) technology is at the forefront in changing the way we live and work.

The use of AI is increasing rapidly in almost every industry, and the world of accounting and finance is no exception. Accounting firms are ramping up their investments in order to automate certain tasks and decision-making typically performed by humans.

AI refers to the simulation of human intelligence in systems or machines that are programmed to think like intelligent beings, mimic their actions and iteratively improve based on the information they collect.

This technology is becoming an essential part of our lives, used in applications varying from digital assistants answering queries online to driverless cars. One of the most visible developments is the use of AI to create ‘chatbots’.

What is a chatbot?

A chatbot – short for ‘chatterbot’ – is a software application that simulates human conversation in natural language through text chats or voice commands. The first known chatbot, a psychotherapist named Eliza, was developed at MIT in the 1960s, inspired by the work of pioneering English mathematician and computer scientist Alan Turing.

The name was chosen, according to the developer, because Eliza’s ability to be ‘incrementally improved’ by various users made it similar to Eliza Doolittle, the working-class heroine in George Bernard Shaw’s play Pygmalion, who was taught to speak with an upper-class accent.

Eliza was programmed to follow the rules and directions of a simple script, simulating conversation by rephrasing statements made by the user as open-ended questions.

This type of chatbot works by identifying keywords from user’s statements and responding according to the pre-programmed rules. For example, if the user says: ‘I am feeling depressed’, the program will pick up ‘depressed’ as a keyword from the statement and respond with a pre-programmed statement associated with the word ‘depressed’, like ‘let’s talk about it, I will listen’.

However, scripted chatbots are limited and rely on a rules-based conversation that follows predetermined paths.

Beyond the script

In contrast, modern, intelligent chatbots are powered by natural language processing (NLP) technologies, allowing them to process and analyse large volumes of natural language data. They can understand the intent of a phrase or sentence, and respond to that intent, rather than to a keyword.

The intent in this case is the task a user wants to be performed or the problem statement a user is looking to solve. This is central to building intelligent conversational interfaces.

For instance, if a user inputs a message like ‘I want to get a tax certificate but I have forgotten my identification number’, an intent classifier would categorise this message as intent to obtain a tax certificate.

Examples of today’s chatbots include Amazon’s Alexa, Apple’s Siri, Google’s Meena, Microsoft’s Xiaoice and Facebook’s M. These chatbots can organise our calendars, play music and movies, order food, control the heating and security in our homes, and even write songs.

Audit is a prime example of disruptive innovation, where RPA is used to automate tasks and streamline processes while increasing speed, accuracy and reliability

The accountant chatbot

Large accounting firms have expanded the range of services on offer by integrating AI capabilities. Many routine tasks can now be automated, allowing resources to focus on more valued-added activities.

Entire end-to-end processes can be carried out with little to no human interaction by leveraging AI and robotic process automation (RPA). Audit is a prime example of disruptive innovation, where RPA is used to automate tasks and streamline processes while increasing speed, accuracy and reliability.

The integration of AI-driven chatbots into accounting software and service offerings is accelerating adoption, enabling business owners to perform standard tasks quicker and more efficiently than before.

Chatbot audit assistants are making audits faster, resolving common questions and removing auditing pain points. They can be used as a customer service tool to help deal with clients, process requests and respond to frequently asked questions – for example, at year end. Payroll accountants can use a chatbot to answer common questions that employees have regarding their compensation.

Accounting firms can deploy chatbots to handle initial inquiries from individuals looking for assistance with tax and other services, converting them from prospects to customers. Chatbots can also be used in cost management to help businesses submit and track expenses and receipts in addition to automating the follow-up process to ensure that invoices are received and processed in a timely fashion.

More human than humans?

Modern chatbots often come with characteristics and personalities in an attempt to make them seem more human-like in the eyes of users. This may include giving them a name and even an age – and even an accompanying background story to help establish a connection.

The addition of a natural sounding voice can build comfort and rapport among users in addition to the obvious benefits of speed and ease of use.

However, there is still a way to go before chatbots are able to achieve the enormous potential they have to create emotional connections with users and deliver a human-like experience.

Most chatbots currently respond to user requests in a formal manner that is often predictable and instantaneous, usually with long messages that overload users with information, leaving them with a feeling that their message was not listened to carefully or understood.

The next step

Future developments are likely to see chatbots that mimic human behaviours more closely, beginning an interaction by engaging users with small talk, taking short pauses between responses and giving succinct replies in a bid to give users the impression that their messages or requests are being listened to.

Many of our devices come with built-in cameras. Facial emotion recognition is made possible by computer vision, enabling chatbots to demonstrate emotional awareness and intelligence. The technology exists for chatbots to model human self-awareness and self-reflection, creating interactions that can make strong, emotional connections with users over time.

These developments will make chatbots more appealing, so that interactions with AI technologies become routine and pave the way for a dramatic proliferation in their use and availability among tomorrow’s consumers.

AI technologies are set to have a significant impact on the digital transformation of accounting and finance. AI-powered chatbots are already in use among progressive accounting firms where menial tasks are automated, freeing up accountants to direct their expertise to add value for clients.

Looking to the future, financial institutions will increasingly use chatbots to converse with customers, provide insights on their spending habits, offer budgeting advice and provide tips that support financial wellbeing.