Dylan Dias | CEO, Neal Analytics
These are exciting times to be in finance, as artificial intelligence, machine learning, automation, and robotics are playing a rapidly increasing role in the field. These technologies enable CFOs to work higher up the value chain, and savvy use of this tech also frees up bandwidth across multiple functions within finance. Three roles – accountant, analyst, and auditor – have seen particularly dramatic shifts, allowing our most valuable asset, our people, to do more of what they do best.
Accountants and bookkeepers have seen one of the most visible shifts, moving from piles of physical paperwork to fully digitized data, resulting in tremendous paperwork reduction, from timesheets to contracts. Being able to move from manually inputting data to scanning and processing saves time and effort, which ultimately leads to cost savings.
In one case I was involved in, a team of four people was doing ad hoc manual data entry on invoices. Using Microsoft’s optical character recognition technology, the key focus was extracting data fields from fax and email documents and having it automatically entered as an electronic record. The initial value-add from automation was about 25%. Over time, this creates a corpus that can be built up as confidence in the solution grows, improving error-handling as they go.
Fuzzy matching is a form of artificial intelligence that compares two sets of data, and allows patterns from high-value documents to be applied to the lower-value documents. This automation enables handling further down the long tail, instead of letting lower-value pieces fall through the cracks. This allows the finance division to cover more ground with fewer people.
An example is the handling of enterprise agreements and incentives with customers. Microsoft had a group of 50 CPAs manually going through these contracts, looking for specific, high-dollar performance guarantees. This solution applied fuzzy matching to surface smaller details from the long tail of the data set that would be prohibitively costly to do manually. It’s a more thorough application of the business rule to the entire corpus, reducing risk. So, it’s a risk play, and it’s a speed and agility play. You can cover more ground with fewer people.
Finance Analyst/ Manager
The biggest change for the finance analyst or finance manager is in forecasting automation. Better forecasts help companies plan and budget better, and save time in doing so. For a large enterprise with multiple lines of businesses, there will be hundreds of people involved in monthly and quarterly business reviews. Crunching these numbers is an opportunity for applying machine learning and analytics to increase the velocity of the forecasting rhythm.
My colleague Jayson Stemmler, a data scientist and project manager notes, “Where we see the biggest impact from a finance forecasting perspective is time savings.” He explains, “One product rolling out now forecasts 700 or 800 different combinations of product segments and geography. To produce individual forecasts for each of those manually would take a large team a number of days – we can do that now in about three hours.”
When we apply machine learning to forecasting, we can help a company understand why it succeeds at its key driver for revenue. Whether the customer is a large beverage company or a nationwide tax preparer, to name two examples, we’ve been able to apply machine learning to sales-driver analysis and found unexpected, actionable patterns influencing growth. We’ve been able to reveal factors the company can control to increase their KPI, whether that’s footfalls to tax offices or cases of soda sold.
In fraud prevention, we find that applying artificial intelligence speeds anomaly detection and strengthens early warning capability. Another core benefit of AI is knowledge mining. Dynamic pattern recognition enables organizations to crunch the numbers of a vast set of data to answer practical key questions. For a philanthropic organization those might include, “Where are the grants doing the most good? What programs have the best compliance?” In work we do with one large philanthropy, we are very aware that there is a vast spread in the personas using the data, from a desk worker in Seattle to a field officer in Kenya. Automation makes sure that each persona has access to the tools they need, in a form factor appropriate to the overall situation.
A key piece of digital transformation
As digital transformation enables these three key roles to generate more and more data, maintaining awareness of it all can become an unexpected challenge. Ensuring visibility across the entire team and keeping data from falling between the cracks will allow everyone in the finance division and the organization as a whole to make better-informed decisions with higher confidence, allowing companies to deliver on their core promises.