Case Studies: AI & data analytics in 2020
While AI and data analytics are very often associated with audit engagements in the accounting profession, the three firms featured below have all found, either by design or accident, the wider applications of the technology in forecasting, planning, consultation and advisory services. And while all three firms featured are midsized or larger, the benefit to small firms is apparent: Analytics is one of the key technologies to making accountants the trusted advisors they strive to be.
Firm: HBK CPAs and Consultants
On the record: Senior manager Antonio Ribeira and principal Christopher Marrie
Product: Mindbridge Ai
Firm size by professionals: 352
Selection: In 2018, Top 100 Firm HBK began an initial relationship with Mindbridge Ai by buying five “uploads.” Uploads refer to one instance of uploading a client’s general ledger into the software; essentially, one audit engagement. Even though this made HBK a very small customer to Mindbridge, the firm was assigned a customer success manager who has always been extremely responsive. This helped convince HBK that Mindbridge Ai would be a good partner going forward, and the firm expanded their relationship.
But it was only after Mindbridge’s user conference in 2019 that senior manager Antonio Ribeira had his “Aha!” moment. “We were spending this money and only using Mindbridge for journal entry testing,” he said. “We were not using the software to its full potential.” He connected with other firms at the conference and put together an implementation process made up of what worked best for several of them. He then went through the firm’s audit process — HBK uses Thomson Reuters’ PPC methodology — and figured out where Mindbridge Ai’s analytics could be applied. He developed a guide and templates for staff to follow — where filters should be applied, for instance, to generate certain reports. Now, the firm is using the software to a fuller potential.
Implementation: In HBK’s early years with Mindbridge, the firm attended webinars on how to use the software. But last year, HBK flew someone from Mindbridge to its Princeton, N.J., office to host a four-hour training.
The most difficult part of using Mindbridge Ai is data ingestion, Ribeira said, which is a theme common to all the data analytics software in these case studies.
“Once the data is in, it’s as intuitive as an iPhone. But if it’s garbage in, it’s garbage out,” Riberia said, referring to client data. “The data has to be configured well and put in correctly. The training covered examples of less-than-perfect client data, and how to fix it.”
Highlights: “I really like their interface,” Ribeira said. “I think it’s really easy to use. It’s really helpful up front when you’re trying to plan an audit because it’s very visual. It’s easier sometimes to look at something than to explain it in accounting terminology.”
He added that Mindbridge Ai has “some great reports you can run out of the software and share with clients. Clients love it; it’s a great value add.”
The more data you can put into the software, the better, Ribeira said. So if HBK can obtain three to four years’ worth of data from a client, they can then run a trend analysis and present a line chart to the client that not only shows the last few years, but can also forecast what the next year might look like.
Until the conference in 2019, HBK was using Mindbridge just for testing journal entries. Now, the firm uses it for planning because of its trend analysis capabilities. HBK principal Christopher Marrie noted that since the implementation of the new revenue recognition standard, ASC 606, Mindbridge AI has been “helpful in auditing revenue transactions, and analyzing them. For instance, if a client has multiple revenue streams, there might be a risk that they could potentially move transactions from a high flow to a less profitable stream to make it look better. Mindbridge can help you find weird transactions like that — things that seem to be anomalies.”
Marrie added that Mindbridge has presented opportunities in consulting, as well: “We had a new client in the $30-$50 million revenue range that had never had financial statements done. The owner said they were thinking of selling part of their company and that the investor asked for their general ledger. They asked if we could look at their GL and see if there was anything weird going on. Without Mindbridge, it would have been too time-consuming, and the answer would have been no. But with Mindbridge, you can find odd entries very quickly.”
In the future, Ribeira sees Mindbridge being part of HBK’s review engagements, and playing a part in due diligence for clients who are investigating acquisitions.
Challenges: “The real challenge is our profession as a whole,” Ribeira said. “One of the biggest issues we have is that the development of technology outpaces our standards. Standards coming out about audit intelligence next year were drafted last year, so by the time they come out they’re based on software two years outdated. The challenge is finding a way to bridge the gap between generally accepted auditing standards and what tech we have today so when someone looks at our file, they know we did it correctly through the software we used.”
Marrie added, “Accountants don’t tend to like change. We tend to not spend time being creative and trying new things, so getting buy-in to a new technology like Mindbridge can be difficult. An audit partner may not want to deal with this in this stage of their career, perhaps. While it might be more work to implement now, in two to three years, it could make your job so much more easy and valuable.”
Ribeira noted that the software hadn’t been peer-reviewed and heavily user-tested yet, so there can be fear from staff around what might happen if there is a peer review of an HBK audit and they get slapped on the wrist. However, thus far, peer review of audits hasn’t posed any problems and there have been no challenges related to the use of Mindbridge Ai.
Firm growth: “In today’s world, a lot of firms are too focused on compliance and not enough on consulting,” Ribeira said. “Our way to differentiate is by using Mindbridge Ai to improve our consulting practice, especially in M&A. That will be one of our firm’s next big niches we launch going forward. Especially coming out of this pandemic, I’m predicting that companies that survive are going to be strong, and those that don’t are going to be gobbled up by the strong survivors.”
Additionally, Ribeira, who heads up HBK’s recruiting efforts for the firm’s four Mid-Atlantic offices, believes that technology like Mindbridge Ai could be attractive to new hires: “Students don’t want to be in traditional accounting anymore. That’s not the vision for the profession. It’s about what type of technology you’re using.”
Firm: Baker Tilly
On record: Principal Dave DuVarney
Product: In-house architected solutions using Microsoft Power BI, Tableau and Qlik.
Firm size by professionals: 2,536
Background: Top 100 Firm Baker Tilly acquired data analytics firm Talavant in March 2020 and folded it into its digital practice. Dave DuVarney, who now serves as a principal at the firm, formerly was president of Talavant, which he co-founded.
Selection: Typically, when a client seeks out Baker Tilly’s analytics services, they are prompted by one of two factors: general lack of visibility into their data, or the need to get data insights on a frequent basis. Especially during the COVID-19 pandemic, DuVarney said, companies need to see their financials in real time instead of 10 days after close.
DuVarney explained that there is a maturity cycle that organizations tend to go through in their analytical journey. They will often start out using Microsoft Excel and doing report-based analytics, which is slow and doesn’t offer a lot of detail. Then they may move to quicker operational reporting, extracting data, putting it into a consumable format, even with MS Power BI, Tableau or Qlik. Then, with the implementation of machine learning, they can try to look forward instead of backward. “There are a lot of solutions on the market that will get people so far, but lead to a dead end,” DuVarney said. “The way an organization looks at metrics may not even be standardized across the business. We help companies understand how they define their data, how they define themselves, and then build to the right level of detail and get the data to the right audiences.”
Implementation: When Baker Tilly takes on a new client for analytics, the firm conducts a series of interviews that cover goals, what they’re trying to do as a business, and what metrics they are using to measure those goals. “We spend a lot of time discussing goals. For instance, if your goal is to increase margin, then you have to measure margin. In a dynamic costing example, you may have to go a step further. Can I look at cost per production run time to understand what it takes to make a product? Are there ways to pull extra margin out by shifting things around?”
The idea is to use interviews to derive, from a client’s objective, what measurable factor Baker Tilly can use to drive to that goal. “We can do all the work, but we tend to be more collaborative,” DuVarney said. “We’re talking about an organization embracing data. We rely a lot on teaming with that org to do it. We want to work ourselves out of a job.”
Highlights: “We would not go to a client and sell them the software,” DuVarney explained. “That’s the wrong approach. We have to understand, at its core, what the strategic value of implementing these things is, then we’ll go find the tool that will help them. There are pluses and minuses to every tool — it’s really about changing the culture of an organization.”
For example, a CFO of one of Baker Tilly’s clients transformed his accounting department from 10 staff who put data together in Excel sheets and emailed them out, to 10 staff who found insights in the information and now help guide business decisions. Another client grew their accounting team from three to 14 people because the value provided by analytics was so apparent to them — and when the pandemic struck their company, they knew exactly what was happening with their business as it was happening, and were able to adjust in real time.
Challenges: “A lot of times, the biggest challenge is the organizations themselves don’t have well-designed business processes.” DuVarney said. “If those are poor, the data they collect is poor. That becomes a consultation opportunity. Now that client moves into Baker Tilly’s transformation, change management, and project management services.” He also said that AI and data analytics are changing all the time, quickly. But at the core, the most important step in deployment of data analytics is understanding what it is you’re trying to measure and why you’re trying to measure it.
Decisions from data
On record: Audit and advisory practice leader John Mark McDougal
Product: In-house architected solutions
Firm size by professionals: 484
Background: Top 100 Firm LBMC acquired data analytics firm Think Data Insights, a provider of data analytics systems and a Microsoft Gold Partner, in 2019.
Selection: Think Data Insights was a natural fit for LBMC, because both firms had strong verticals in health care, manufacturing, distribution and private equity — and it’s in private equity that the firm really sees the power of analytics-driven reporting, McDougal said.
“We’re there to help clients obtain, view, report, disseminate, automate, make decisions from data, look for trends, develop machine methods of looking for those trends, and reporting back to the people to make decisions,” McDougal said.
Implementation: For clients that need data analytics, LBMC can offer either a pre-packaged solution, or a custom-architected solution. The firm also develops apps in areas like reporting, and the tracking of information such as internet sales for a retailer.
Highlights: The biggest selling points for AI-driven data analytics, McDougal said, are consistency in reporting and commonality — common platforms, and common language for discussion around data.
LBMC offers a service called Dashboard in a Day, which is a shorter-term project for clients that need a less intensive analytics engagement. “Analytics can help move even just one small aspect of the business forward, and that can make a difference,” McDougal said. “Dashboard in a Day is a series where we help clients put their own dashboard together or get training to do it themselves. This has been useful for lower or middle-market companies in the $5-$50 million range. We’ve seen private equity firms use it too.”
Challenges: As noted by the other firms in this case study, data ingestion is the most common challenge when conducting analytics. But also, understanding client context for that data is important. “You have to understand the data from your client’s point of view, because sometimes what they call something and what they believe it is, is not what we would call it. We listen and listen and listen to our clients,” McDougal said.
For instance, a simple phrase like “repeat customer” could mean very different things to different businesses. The parameters are important to understand, so that when mining data and creating reports, the firm understands what the client needs from that data.
“We take a holistic approach with our clients,” McDougal said. “Being recognized experts in the field of analytics, we get the call to work in that realm with our clients — but we are routinely in business and advisory conversations. We are sharing not so much just the service, but talking to them about what they do in analytics and AI currently. And sometimes it’s a lot, and sometimes that’s the first time they’ve had the conversation.”
Opportunities for small firms: AI-driven data analytics seems like leading-edge technology beneficial mostly for large companies and by extension the large firms that serve them. But McDougal believes there is opportunity for smaller firms to deliver these capabilities to their clients, too. “I’m thoroughly convinced AI data analytics is not specific to larger firms or larger clients,” McDougal said. “It’s something that, where applicable, can help any nonprofit, for-profit, any business — and any firm has the ability to offer value.”
It’s obvious, McDougal said, that a company like Coca-Cola, for instance, has a lot of data to use to spot trends. But a $20 million-a-year internet retailer going B2C has trends in customer buying patterns, site traffic, or customer base preferences that can be exploited to ultimately increase net revenue per order and achieve natural growth year-on-year by a few extra percentage points, if analytics are applied strategically.
“Analytics can move an organization over a hump and toward improvement and increases, particularly related to sales,” McDougal said. “And lack of analytical ability, or the lack of ability to gain insights into data, has been an impediment toward getting some routine business decisions made. It’s been a stumbling block.”