Machine learning and AI are terms which get thrown around often. Whether or not you’ve bought into the hype, one thing is for sure, it’s something rarely discussed in much depth. Jahan Zahid, CEO of Indigo cashflow, held a panel discussion on “Machine Learning, AI and its Impact on Accounting” to try and address this.
Pictured above: (Left to right) Daniel Houghton (Tide Bank), Damon Anderson (Xero), Wincie Wong (RBS), and Adam Goodall (Coconut).
Reconciling Bank Transactions
Xero’s recent Digital or Die Report reveals that 61% of accountants spend most of their time reconciling bank transactions. We asked the panel how they see that changing.
Adam Goodall, co-founder of Coconut, sees it as an upstream data problem and more about “how the data gets into the accounting package” in the first place. For him it’s about keeping payment and receipt together at the point of transaction.
Wincie Wong, digital propositions lead at RBS, shared their plans to solve the receipt capture problem through their partnership with Sensibill beginning early next year on their mobile banking app. The pain point they receive from their customers is at a very basic level of “knowing where your missing receipts are”.
Damon Anderson, head of UK partnerships at Xero, believes there’s a huge opportunity with machine learning and Making Tax Digital (MTD). For him a big part of their strategy is removing the manual entry and compliance elements and “repositioning the role of the accountant from the back office to the front office”. Their vision is centred around “code free accounting” and removing any of the traditional accounting from the customer.
Daniel Houghton, who heads up partnerships at Tide Bank, believes that if firms don’t adapt digitally they end up “bogged down in admin” and get “killed by being left behind”. For him, not having clean data feeds stops businesses from being able to make those key decisions which lead to the “1% differences”.
Approaching MTD Deadline
The exchequer will surely be looking forward to MTD coming into force, hoping to address the £3.5bn shortfall in 2014/15 revenues, due to mistakes in VAT returns alone. We explored what’s likely to happen as we approach the MTD deadline.
Anderson sees the compliance element as “way ahead of the game” and would like to see a future where accountants more readily adopt the “advisory stuff, in predicting where a business is going to be”. He sees the opportunity for machine learning to come into such services like cash flow forecasting.
Will compliance for the accountant be going away anytime soon? Not likely, according to Goodall who highlights the fact that “there’s differences between businesses that you can’t always account for, qualitative factors that aren’t really feeding into those models”.
Anoop Rehal a Director from BDO UK LLP, who was at the event, agrees that the advisory piece is important but worries whether businesses will get the right advice. Part of that for him stems from accountants needing to be taught how to be advisors in the first place.
Will accountants need to learn to code?
With a danger of completely relying on machines to make decisions, we discussed whether accountants will ever need to get under the hood and understand computer code.
One of the audience participants was quick to stress the importance for accountants to “understand how information is generated in the first place”, and felt nervous in an environment which relied completely on the machine.
Goodall believes it’s going that direction, especially with audits likely to become more technically demanding.
Houghton was less sure about the coding aspect, but sees the accountant needing to be “absolute experts, about the software, and how it works”. He also foresees having potentially different software versions depending on the interpretation of the law you’re comfortable with.
Wong thinks "it depends on the kind of accountant you want to be, there’s always going to be people who are going to be more technical”.
It would be interesting to hear what readers think about this.
Predictions for 2018
- Goodall: We may see the first legitimate use of blockchain in the space
- Wong: With Open Banking coming into force early next year, we can expect to see innovations in applying that to create new digital experiences
- Anderson: Removal of early processes at the point of sale, before transaction data gets into the accounting system
- Houghton: Google and Amazon providing more coding as part of the transaction and for that to show up on the bank feed
About Jahan Zahid
Jahan is the founder and CEO of Indigo Cashflow, an app created to take the pain out of cash flow forecasting. Prior to this he ran a successful software consultancy, developed FX trading models at Bank of America Merrill Lynch using machine learning techniques, and worked as a Mathematics Lecturer at Bristol University. He holds a PhD in Mathematics from Oxford University.