I spent some time last week in Atlanta at a tax technology forum courtesy of our good friends Tim Howe and Nick Patel from Synexus Tax Solutions. Originally, I thought this forum was going to cover all things sales tax tech, but I ended up being pleasantly surprised as we quickly spiraled into what the accounting firm of the future will look like from a technology standpoint (there was some sales tax talk). After the enlightening day of plenty of tech-speak that went over my analytical brain, I frantically deleted my original blog post about taxpayer-funded stadiums. We’ll save that one for another day. I had found a new, timely topic. Anyway, after eight hours of discussion, the day can be summed up into one word: automation.
It’s no secret that the number of new accountants entering the profession each year is shrinking, but that doesn’t mean the profession is going away. Quite the opposite, actually. It’s just being reimagined. Instead, we are adapting to find ways to make menial – but absolutely necessary – tasks more automated so that we can have time to focus on value-added client service. Dashboarding to track project completion progress, current status and deadlines at the click of a button seems to be table stakes these days. Automating engagement letters and return form population are trending in that direction as well. The logical next step (if it’s not here already) is artificial intelligence (AI) and machine learning, essentially teaching themselves how to read tax/audit documentation and either make taxability determinations, identify accounting errors, or even provide research and responses to controversies. The amount of time and effort that is being invested in this “new wave” of accounting automation is quite staggering. The speed at which many of these programs are being developed is similarly jaw-dropping.
This isn’t to say the bots are taking over… yet. Just like everything else, these items are meant to be a tool to help us do our jobs more efficiently, not become a full replacement. While automation is great, plenty of wrecked self-driving cars are in the boneyard. Who here hasn’t used their favorite AI platform to fashion a past-due article or an email or at least entertained the thought of doing so, only to discover that the result is only a crayon away from being completed by a kindergartner?
We’ve discovered very quickly that there are flaws in many of these systems when we don’t live in a static tax world, especially when human interpretation is involved. Letting AI make a tax determination based on whether your “to go” meal was served with utensils or whether or not attendance at a trade show that included a prospect meeting over cocktails but where no actual sale was made is enough to create income tax nexus can be a slippery slope. Taxing authorities, after all, have been known to contradict themselves a time or two. These programs are designed to provide a response even if it hasn’t had a chance to “learn” the correct response. The result is a so-called “hallucination” effect that returns a response out of thin air. Hopefully, you haven’t fallen victim to a citation that didn’t exist.
We are at a point where we, as accounting professionals, have access to many tools that will make our lives easier. The key is that they must remain tools and not become complete replacements. So keep the process automation and efficiency coming, but I’ll write my own appeals… for now.