How we successfully implemented AI in audit

A partner at Garbelman Winslow CPAs describes the small Maryland firm’s experience using machine learning to improve its audit process.
By Samantha Bowling, CPA, CGMA, as told to Cheryl Meyer

 Samantha Bowling, CPA, CGMA, a partner with Garbelman Winslow CPAs, says using artificial intelligence in audits is more comprehensive than random sampling.
Samantha Bowling, CPA, CGMA, a partner with Garbelman Winslow CPAs, says using artificial intelligence in audits is more comprehensive than random sampling. (Photo by Ariel Lewis/Bonnie Johnson Photography)

Garbelman Winslow CPAs is evidence that you don't have to be a large firm to have big aspirations with cutting-edge technology. The Upper Marlboro, Md.-based practice, which has a total of six CPAs and 15 employees, is using artificial intelligence (AI) to identify high-risk transactions as part of its auditing process. Here's what the firm has done and learned so far, as told by partner Samantha Bowling, CPA, CGMA.

We were innovators back when computers and mainframes were introduced. Today, all of our 15 employees have desktop and laptop computers, monitors, and virtual work-from-home access. We're paperless. The only file cabinets here are from a firm we acquired, because we have to retain those files for seven years. And we just moved our in-house server and phone system to the cloud within the last year. We used to have seven servers on-site. Now everything's in the cloud. We used to be able to remote in to our servers before, but it was slow. The remote access is much better in the cloud. It's night and day. I'm excited because I've never had access to this much technology.

We recently implemented AI for auditing. The AI we use is machine learning where the machine has built-in algorithms that help it learn based on transactions it is fed. This software platform uses AI and various control points (benchmarks) to analyze the transactions in the general ledger and then puts those transactions into buckets: high risk, medium risk, or low risk. Previously, we would manually sample at random to figure out which transactions to review, which was less effective. AI is absolutely more comprehensive; it alerts us when things don't look right and tells us where to start and where our risk is going to be. It shows the risk at the transaction level.

I started looking into AI several years ago when I was on the AICPA governing Council. AICPA President and CEO Barry Melancon, CPA, CGMA, described how the Big Four firms were spending millions of dollars on AI. I had this fear that if our firm didn't invest in AI, we'd lose our auditing business. AI would become a larger-firm technology. This is a huge opportunity for small firms: We don't have to lose our competitive advantage, and we're not stuck on legacy platforms.

For over two years I searched for a suitable AI platform. Some we could not afford. Then, in 2017, I found a Canadian company called MindBridge Analytics Inc., which had cloud-based platforms for small, medium, and large firms, based on the number of users and clients. I called them and said I would like to work with this software and give them my feedback as a small business owner who does audits. For a discounted price when I first signed, I was given unlimited access and users in return for my feedback.

There were a few hiccups in the beginning. I worked with MindBridge for about six months, giving continual feedback on what they needed to change and add, since Canada uses different terminology and audit processes. We had an initial training session, and it was user-friendly. MindBridge gave us a login to its application and secure portal. I sent them general ledgers, and it took two weeks to get access to the risk report on the client dashboard. Today, I have direct links to the software we use and can get a report in 10 minutes. It's amazing. The great thing about working with a small software company is they are willing to listen and make changes. Small public accounting firms can implement this technology for less than $10,000. The price depends on the size of the audit practice, but even if you want to try it out with one client, it's worth getting your feet wet.

We now have a competitive advantage. I work with not-for-profits that use QuickBooks Online and other cloud-based software. I go into the cloud for MindBridge, click the QuickBooks link for our clients, and link my clients to the platform. Then it pulls in all the information. I'm now able to more accurately pick a relevant sample for my audits based on risks at the transactions level. This should help us grow our audit practice. This gives us a competitive advantage over firms that use the traditional statistical sampling. We are selecting our sample by using AI to review all transactions and pick the riskiest transactions for our sample.

We also know what fees to assess our clients. Usually when we bid for an audit, we are never sure what we're getting ourselves into. So, we assess a range of fees, and a lot of times we end up losing money because the first year is a mess. If I get a potential new client today and I haven't bid yet, I put their data in my audit platform for AI and see what it tells me in terms of risk. If I see a lot of high-risk transactions, then I can either not accept the engagement or price it appropriately based on the risk level. We've never had the ability to do that before. If a firm won't share its data, I turn down the engagement. I will not work on an audit without putting the financials in the AI software first. It is an audit, and we have the tools to make educated decisions about which clients we want to work with.

We think AI is a great tool for controllers and internal audit departments. Controllers don't have time to look at all transactions posted by their staff, but I bet they would like to look at the high- and medium-risk transactions. Also, I am pretty sure the internal audit department wants to address the risk issues before the external auditors inquire about them.

Some CPAs think their technology processes and systems are not broken, but they are. They just don't realize it. So, they need to ask, "What drives me crazy about what I do every day, and how can technology fix that for me?" And I guarantee there's a solution. CPAs should get involved with their state CPA society or the AICPA on a national level to learn more about technology and how embracing it makes them better CPAs.

We're in an innovation revolution. The change of technology is so fast that we can't ignore it. Many industries are being disrupted by technology — such as Uber taking business from cabs — and nothing is stopping anyone from saying, "Hey, we can do it better."

We are also using AI for risk assessment and looking for ways to market this to nonaudit clients. I expect that we will be using new technologies in all the services we provide. It's more of a total audit, and it's going to change the way we do auditing. The possibilities are endless, and I am excited to see what is around the corner.

Cheryl Meyer is a freelance writer based in California. To comment on this article or to suggest an idea for another article, contact Jeff Drew, a JofA senior editor, at

See Part 1 of the JofA's annual technology roundtable, "What's 'Critical' for CPAs to Learn in an AI-Powered World," as our panel of technology experts discusses the incredible potential of machine-learning technologies and how accounting firms need to adapt their thinking and update their skill sets to benefit from these advances.

Where to find February’s flipbook issue

The Journal of Accountancy is now completely digital. 





Get Clients Ready for Tax Season

This comprehensive report looks at the changes to the child tax credit, earned income tax credit, and child and dependent care credit caused by the expiration of provisions in the American Rescue Plan Act; the ability e-file more returns in the Form 1040 series; automobile mileage deductions; the alternative minimum tax; gift tax exemptions; strategies for accelerating or postponing income and deductions; and retirement and estate planning.