The adoption of artificial intelligence systems in the world of corporate finance has many benefits, among them the ability to spot issues before they cost the company money.
The critical factor in using AI is acting with purpose when presented with data, which brings up an interesting dilemma: when is the best time to act when issues arise?
Yes, AI can identify problems in “real time” but that doesn’t mean action should happen right away. It’s often best to act at the “right time” when confronting problems.
Right Time vs. Real Time
Let’s look at this real time vs. right time discussion in another context: healthcare.
For example, if you have a dramatic or sudden health problem—a stroke or a gunshot wound, for example—you need to get to the emergency room right away. To save your life, doctors must treat you immediately, so real time action is best.
But if you have a sore throat, you shouldn’t go to the ER because it’s more efficient and cost-effective to see your family doctor. In this case, the “right time” approach is best.
Fortunately, we have very few gunshot wounds in corporate finance.
If your company has an AI-based robo-auditor system that spots a fraudulent wire transfer, you should take immediate action to stop that transfer.
But for most problems, real time is not the right time to act. It’s better to study what your analysis shows and determine actions you must take to correct the problems.
A great example of “right time” action is addressing wasteful spending by employees.
When a robo-auditor flags the expense report from an account executive who lavishly overspent on a trip to New York, the temptation is to come down on him like a ton of bricks. Those decisions usually hurt morale and don’t encourage positive change.
The better approach is to use the robo-auditor to train that account executive on how to make smarter decisions in the future. Those suboptimal spending decisions can be used to explain how the executive’s travel budget would be affected if those costly mistakes continued. That’s a “right-time” decision that gives you a better long-term result.
Patterns Driving Process Improvements
Most opportunities to improve will come from influencing future activities, not just from correcting a particular problematic transaction. One of the biggest benefits of AI systems is that they can be used to show people the net effect of small decisions.
When business travelers avoid expensive hotels, it stretches their travel budgets, so they can afford to make more trips for the same money. More trips mean more revenue for the company and higher commissions for the executive. Everyone wins.
The “right time” depends on what you try to influence, control, or fix. If you use AI to drive an autonomous car, you want to instantaneously correct course to avoid a tree.
But if you seek to prevent duplicate payments, you need to check things as frequently as you make payments—whether it’s daily, weekly, or whatever is appropriate.
You don’t want to change corporate policy concerning travel every five minutes, but it might make sense to re-examine it every quarter.
For any kind of process improvements, it’s important to see patterns and not just individual events, and to see patterns, you need to look at things over time.
By being patient and using the data an AI system brings you, it’s possible to correct behaviors, not just specific issues. A one-off approach might recover money in the short term but using education to course correct pays bigger dividends over the long haul.
This is the second in an ongoing blog series based on the recently released book, Robo Auditing: Using Artificial Intelligence to Optimize Corporate Finance Processes by Patrick J.D. Taylor, Manish Singh and Nathanael L’Heureux.