<img src="https://ws.zoominfo.com/pixel/BUJfPb8NrEnpjSiz8kRz" width="1" height="1" style="display: none;">
Company Culture Artificial Intelligence

How to Get Started with AI and Robo-Auditing

on October 10, 2018

If you’re a CFO who’s decided to implement a robo-auditing system within your company, the question you’re probably asking yourself is: how do I get started?

Artificial intelligence (AI) is undoubtedly complex. But the process of setting up an AI system and training employees on how to use it doesn’t have to be.

CFOs who adopt a robo-auditor should expect their team to be up to speed with the system in six to nine months. At this point, a formal review should be conducted.

What should happen in those first few months to ensure employees hit benchmarks during that review? You’ll need a plan to make sure nobody gets lost in the weeds.

You’ll also need a long-term plan to ensure the AI system’s potential isn’t wasted after it’s set up and begins to learn the nuances of your company’s systems.

In this article, we’ll lay out a guide to getting started with a robo-auditor: where to begin, what habits to change, and ways that AI can help improve your processes.

Train Your Team in Chunks During a Phased Rollout

The first step is training employees on how to use a robo-auditing system.

In our experience, a one- or two-week course on how to use the new system doesn’t lead to good retention. People go numb drinking from a fire hose that way.

We’ve found it to be more effective to provide smaller bits of training followed by hands-on practice working with the system in a reliable, high-impact area.

The first step is finding a problem and assigning the resolution to someone. Start with black-and-white cases like duplicate payments that are easily figured out.

You want employees to learn the basics, then come back a week later for more training to implement the next capability. You can also employ online, self-paced training to do this. Employees who put the training to work right away cement the knowledge.

The second step is tackling issues that fall into more of a gray area. Context comes into play as employees seek to understand the “what” and “why” of the robo-auditor’s data.

If the flagged issue is complicated, like a CEO’s travel expenses, they should reach out to someone else in the organization to see how to react in future instances.

Once the issue is carried through the resolution process, the AI system will learn from that action and not flag the CEO’s large travel expenses in the future.

You want to go after highly reliable things, so you can demonstrate early results. Don’t try to do the hardest things first. Crawl first, then walk, before you try to run.

Your primary goal for the first six months should be to make steady improvement. A phased approach gets easy wins that can be expanded upon moving forward.

Be Prepared to Let Go of Unnecessary Habits

AI and automation dramatically reduce the time needed for the audit process. It should be at least twice as fast as traditional auditing, and more thorough.

As your system gets up to speed, be prepared to let go of old, ineffective habits. For instance, you may no longer need to have auditors review expense reports.

It’s common for the accounting department to examine a random sample of 20% of the company’s expense reports to make sure the rules are followed.

The incidence of problematic expense reports is low (typically less than 5%). Because of this, the time spent looking at the sample is a waste of the company’s time and money.

With the robo-auditing system, you don’t need to spot check anymore because the robot looks at everything. It finds only those expense reports that are different or have an anomaly.

These reports don’t always have a problem, but they are all interesting or different, meaning they’re worth review from a human auditor. Now your human auditor only looks at about 3-5% of expense reports instead of 20%, saving time and money.

With oversight like this, many organizations eliminate manager approval of expense reports. The idea is shocking at first, but in practice, it removes an ineffective step.

Most companies use automated expense report systems that make it easy for the manager to hit the approve button. A manager may be diligent, but reviewing expense reports is time-consuming, so the review usually isn’t rigorous.

Furthermore, a manager who doesn’t review reports right away delays reimbursements, and those delays are a source of friction with employees.

Doesn’t it make more sense to eliminate that step altogether?

Instead, a robo-auditor can flag different types of policy violations and send a dossier to managers for them to review patterns of wasteful behavior with employees.

Managers don’t get bogged down with individual expenses because they’re having more substantive conversations that will influence future employee behavior.

Look for Ways that AI Can Help Improve Your Processes

In addition to automating auditing work, an AI system allows you to leverage computers to keep up with details that may not have been tracked before.

Those details come from automating the analytics (what happened) and the resolution process (how it was fixed) and recording what happened. In doing this, the robo-auditor creates a data set with which to improve fundamental business processes.

Those can be big or small changes, but by keeping up with the aggregate results of the analysis in a disciplined manner, you can have a real impact.

Data is only useful when it’s followed up with action, which is why it’s important for companies to utilize a system with controls for both analytics and resolution.

Well-designed systems will require that you assign a reason why something happened to put the exceptions in categories. This allows you to go back and see if you have instances piling up in a particular category and look for ways to fix that issue.

In this way, the robo-auditor tracks why problems occurred and aggregates them. This kind of root-cause analysis allows you to redesign things to eliminate the problem.

In companies without an AI system in place, this process isn’t as effective.

For example, when companies manually send emails or call employees about problems with expense reports, they may not track whom they’ve emailed the most, which robs them of the chance to have a discussion that could drive long-term improvement.

One company we worked with saw that multiple employees were using corporate credit cards for personal purchases, and when they asked why, it was discovered that these employees were confusing their corporate and personal cards.

Whether that reason was legit or not, the company issued new cards with a prominent company logo on the front, thereby eliminating that excuse for corporate card misuse.

Now when their cards are misused, the company can offer counseling to an employee who’s in dire financial straits or fire ones who continually commit fraud.

Many things in the business world are in black and white. But some of it is in the gray area, too. You’ll never eliminate that gray area, but AI helps you shrink the gray.

An AI system can also save you time by learning the nuances of accounts receivable.

Say your corporate agreement dictates that a customer’s payment is due in 45 days, but one of your long-standing customers never pays before 60 days.

The AI system can recognize this pattern of payment in 60 days and save your company the time and expense of sending out reminders after 45 days.

This also saves your dependable customers the aggravation of getting bills they aren’t ready to pay yet, plus the AI system can alert you when it has been 61 days.

This is the eighth blog in an ongoing 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.



Manish Singh

Manish Singh is Executive Vice President of Sales and Client Success at Oversight Systems.