AI will certainly have an impact on the way that retail accounting work is done, but there is still debate as to the depth and timing of that impact
Artificial intelligence (AI) is all everyone in the business world seems to be talking about today, everything from how it will be used to improve business processes, and which kinds of employees it will replace. And many corporate chief financial officers (CFOs) already are evaluating the potential of AI as a way to increase profits while reducing costs.
Further, a new generation of chatbots is taking over online customer service for businesses, while many firms are using ChatGPT instead of freelance writers to develop content for blogs, websites, email campaigns, and social media posts. Others are using AI to analyze customers’ online activities to better boost the effectiveness of targeted email marketing campaigns.
AI is also beginning to make baby steps into corporate accounting departments. A new generation of accounts payable and payroll platforms uses rudimentary AI to read scanned invoices, bills, payroll records, 1099 and W-2 forms, and tax notices and then automatically categorize and enter this information into the department’s accounting software.
While these uses of AI can help free up time for accountants and bookkeepers who once had to enter this information manually, there is one critical area of accounting where AI is unlikely to make headway soon — account reconciliation for retailers and restaurants that generate significant revenue through online sales.
And that’s unfortunate, because in this era of razor-thin margins CFOs need to have access to accurate profit & loss information every day. The accountants and bookkeepers they rely on to deliver this information, however, are getting so frustrated by what has become a very labor-intensive and time-consuming process that many are quitting in droves — and CFOs are having a very difficult time replacing them.
What AI could do… and why it can’t right now
In an ideal world, an AI-based account-reconciliation platform would automatically download all sales and payment data from every system used by every location, analyze and categorize adjustments, flag exceptions, and prepare summaries for the accountant to review and approve before entering this information into the company’s accounting system.
Unfortunately, this isn’t going to happen anytime soon. And here’s why.
Not only is gaining access to sales data difficult, but an even harder part is making sense of the data. The ability of an AI system to learn is completely dependent on the quality and consistency of the information that is fed into it. AI works best when it’s trained on structured data that follow predefined rules.
A new generation of accounts payable and payroll platforms uses rudimentary AI to read scanned invoices, bills, payroll records, 1099 and W-2 forms, and tax notices and then automatically categorize and enter this information into the department’s accounting software.
This isn’t an issue for payroll records, tax forms, bills, and invoices, which tend to use relatively consistent data structures that make it fairly easy for AI systems to learn and process. Unfortunately, no such consistency exists in the ecommerce world. Each vendor uses siloed sales data formats, with widely variable and often-opaque structures for reporting adjustments. Indeed, some even use their own custom-built platforms with no proper reporting for accountants.
Take state sales taxes, for example. Platforms like Shopify automatically pay taxes on sales made through Instagram and Facebook Marketplace, bt Shopify doesn’t clearly label tax adjustments in its transaction records. Instead, it’s up to the retailer’s accountant to identify sales taxes and countless other non-annotated adjustments in transaction records, make sure that net sales and deposits match, and resolve exceptions when they don’t.
Every time a retailer signs up for a new ecommerce platform or service provider, this creates yet another sales data decoding challenge for accountants and bookkeepers who are already struggling to master existing platforms.
An attrition risk factor that’s not going away
Unfortunately, CFOs can’t afford to ignore these problems, because burnout among bookkeepers is becoming a key challenge for retailers.
Further, this is a relatively new phenomenon. Before sales apps like Stripe and ShopKeep launched in the early 2000s, retail account reconciliation was relatively easy. Most brick-and-mortar retailers generated all of their sales on-site and most accountants and bookkeepers only had to reconcile cash and credit card transactions and download sales reports from a single point-of-sale system.
In recent years, however, everything has changed. Retailers are using upwards of five sales and payment platforms without scrutinizing the quality of the sales reports generated by the systems. As a result of these decisions, many retailers are paying the price in terms of increased attrition among their accountants and bookkeepers.
How bad is the problem? Between 2000 and 2022, more than 300,000 accountants quit their jobs, and today the average turnover rate among corporate accountants is 13.4%. And the outlook for bookkeepers isn’t much better. Job growth is expected to decline by 5% through 2030. More problematic is that experienced and knowledgeable veterans are leaving the industry in increasing numbers. As a result, today the average bookkeeper has less than two years of experience.
What CFOs can do in the meantime
Unfortunately, a universal ChatGPT-style AI solution for retail accounting isn’t on the immediate horizon. Not as long as the tax & accounting industry doesn’t come together to develop standardized data structure conventions.
CFOs who aren’t quite ready to partner with outside accounting automation firms can still support their accountants and bookkeepers by giving them a voice when it comes to choosing or changing sales platforms.
However, progress is proceeding on a decentralized scale. Accounting software developers are beginning to move accountants towards data workflows that can convert siloed data-reporting formats into digestible sales data that can eventually train AI systems to process them automatically.
And some of these developers already offer solutions that can automate account reconciliation for retailers using many online sales platforms. For example, automated accounting firms like Bench and Pilot can automatically collect, analyze, and reconcile sales data from systems like Square and Shopify for their accounting clients.
In the meantime, CFOs who aren’t quite ready to partner with outside accounting automation firms can still support their accountants and bookkeepers by giving them a voice when it comes to choosing or changing sales platforms.
Vendors should be willing to provide samples of downloaded sales reports and transaction records they can review by accountants to assess how much effort it could take to decipher them. And accounting professionals should have the opportunity to meet with members of the vendor’s development team to ask questions and express concerns about any AI-driven software solution.
CFOs also should be willing to reject any vendor whose sales feeds would create additional and unnecessary work for their accountants and bookkeepers. Or, at the very least, press vendors to explain how they are improving the company’s financial reporting and whether the sales data can easily be processed by accounting automation platforms, in case the retailer chooses to partner with one of these firms down the line.
Keep in mind, the easy part is adding new ecommerce platforms to your business. The more challenging aspect is finding and retaining tax & accounting professionals who are willing to work with sub-optimal systems. This is one situation where the value of human capital is far more important than simply making more sales.
The author is co-founder and CEO of Bookkeep, an accounting automation platform that works with existing finance teams to help them normalize and automatically post accounting entries from over 50 plus ecommerce and point of sale platforms.
