The Modern Finance Forum’s most recent research reveals a healthy degree of scepticism about the prospects for artificial intelligence (AI) and machine learning in the finance function. In a September 2017 survey of more than 850 senior finance professionals, 56% of respondents expected to become more dependent on machine learning to drive more accurate forecasts, but only a quarter (24%) felt that machines will be better at predicting the future than humans!
It seems that the finance profession is not ready to relinquish its role to accounting robots. They don’t believe that professional judgement, life experience and even ‘gut-feel’ can be reduced to a few lines of programme code. But that doesn’t mean that there isn’t space for AI in the finance function, and here’s why.
1. We already let machines make decisions for us
It might not be at the front of our minds, but we are surrounded by learning systems that influence our lives from the moment we rise from our beds. The “Nest” heating App has movement sensors that learn users’ behaviour and it ‘knows’ several weeks ahead when it should switch the heating on and when extra hot water is needed. The Waze navigation system tells users the most efficient route to work based on current traffic patterns, and Amazon suggests which products on-line shoppers might like to buy based on their e-commerce journey through the site and browsing history. So why should ERP systems be regarded as lifeless and inert? Surely, there must be decisions the system could make on its own?
2. We have a higher level of trust in our ERP data
ERP systems is the finance function’s primary source of data and most of the information entered is subjected to significant controls and validation. By contrast business information systems like corporate performance management (CPM) software, draw their entries only partly from ERP, incorporating a wide range of less codified data sources and spreadsheets as well. Many of these have not been subject to the same accounting rigour, are more scattered and fragmented, and thus less trustworthy. Finance professionals looking to drive additional insights through AI may find it more challenging to start their journey with their CPM systems.
3. Unified ERP provides a well-defined data ‘universe’ for driving insight
The challenge for AI and machine learning is that its effectiveness is extremely dependent on good data. ERP frequently has a rich source of trustworthy data (financial and non-financial) on hand, especially when it is integrated with other systems like Customer Relationship Management (CRM), Human Capital Management (HCM) and supply chain. Historically, these data sources have been under exploited or even ignored by the finance function, but the tide is turning. Our recent research shows that 26% of CFOs rank the CRM database as their most insightful source of data just behind the general ledger with 34%.
Imagine the benefit of visibility into customer behaviour data, and the ability to analyse it in novel ways. You could identify and prevent customer churn, spot their propensity to buy goods and services and use this data to increase sales, and objectively (statistically) validate buying patterns to inform more accurate revenue forecasts – rather than relying on sales managers’ quarterly forecasts.
The announcement last year of “Einstein for Salesforce” (sure to be followed by more AI inspired initiatives) is profoundly game-changing for the finance function, as it dips its toes tentatively into the AI space for the very first time. ERP is the low-hanging fruit that allows modern finance professionals to drive insight from data they already own and trust, in novel and imaginative ways. AI for ERP will be the springboard for a whole new generation of capability.