FSN research identifies that 52% of finance professionals spend too much time on transaction processing; 42% too much time on regular management reporting and 32% say they are spending too much time on statutory reporting. It appears that CFOs and finance professionals are overwhelmed by the basics of accounting, leaving precious little time to devote to their broader remit of being a business partner, advisor and strategist to the rest of the management team.
The simple application of the Pareto principle or the 80:20 rule as it is commonly known, suggests that 80 percent of transactions will be processed correctly first time and 20 percent will fail for some reason or another. The trouble is that an errant transaction costs 80% more to process than a transaction that falls perfectly into place first time, and around a quarter of everyone’s time is spent daily simply searching for information that will help resolve some issue or another. The maths is compelling. There is every incentive (economic, operational and technical) to whittle down the impact of anomalous transactions. And with data volumes virtually doubling every two years the challenge is becoming more pressing by the day.
So the obvious question is, how can finance professionals tip the balance in favour of straight-through-processing?
Innovative technologies are beginning to make an impact, especially anomaly detection powered by artificial intelligence (AI) and virtual, or so-called digital assistants. As with all instances of the pareto principle, the conundrum is that we know that 20 percent of the transactions need further scrutiny, but we usually don’t which 20 percent.
It is also noteworthy that an anomaly does not always represent an error or a risk. For example, anomalous variances and outliers could point to an opportunity that nobody has discovered or exploited. Anomaly detection therefore opens up new vistas of opportunity as well as trapping risks before they become a problem.
AI utilises the power of the computer to a large body of transactions in a matter of seconds – something that would be impossible for human beings. Finance professionals can define the business rules and parameters for anomalies, for example, what characterizes a risky transaction and then leave the machine to do the rest. A risk-based approach quickly identifies the highest risk transactions that require further investigation, but more importantly, with aid of machine learning, allows finance professionals to progressively reduce the level of anomalies. For example, it is not unusual for reconciliation errors to be reduced eventually to under five percent using anomaly detection, aided by machine learning.
Virtual finance assistants take automation a stage further by enabling the monitoring process to be fully automated. Rather than a human invoking anomaly detection a machine can quietly run anomaly detection routines in background and serve up the reports to the right people at the right time via workflow. So called, “lights-out-processing” enabled by virtual assistants massively extends productivity by allowing computers to continue working throughout the night. Furthermore, Natural Language Processing, i.e. the ability to invoke processes such as report production or anomaly reporting with voice commands, allows a more intuitive method of working with advanced technologies that further increase productivity
But enlightened software vendors such as Prophix, know that savvy CFOs do not want to employ legions of data scientists and other specialists to develop AI applications. The key for them is that AI capability, machine learning and virtual assistants are embedded directly into the applications to make them natively more intelligent and accessible straight ‘out of the box’. Prophix’s “Anomaly Detection” and “Virtual Financial Analyst” do just that.