AI, ETL Tools Offer Promising Possibilities

print this article

Artificial intelligence (AI) and extract, transfer, and load (ETL) tools are transforming what many TEI members do every day. We wanted to find out how these dynamic technologies will impact the tax industry. In May, Michael Levin-Epstein, Tax Executive’s senior editor, interviewed Nick Panko, vice president at CFO Services. 

Michael Levin-Epstein: Nick, tell us a little bit about CFO Services and what you do there.

Nick Panko: CFO Services, we’re a boutique tax consulting firm. We focus on credits and incentives, specifically a few things: federal incentives [like] R&D credits, and energy credits like the Inflation Reduction Act credits, and then CHIPS Act credits as well. We also focus on state incentives; every state’s got jobs and investments, so when our clients are expanding out or hiring, investing in capital, we can identify those programs and help them keep track of that long-term compliance. Then, third, we provide technology services, both in our incentives areas but also in the Microsoft stack, with technology consulting for tax departments on how to make their processes better and automate tax processes.

Levin-Epstein: Could you tell us how tax departments can start implementing extract, transfer, and load (ETL) tools and take advantage of AI capabilities?

Panko: Our work in tax technology, again, is helping those tax departments look at their processes, where can we streamline and make more efficient. And in terms of ETL tools—extract, transform, load—a lot of times during either provision, planning, or compliance, our tax departments are spending a lot of time massaging data or transforming it into the analysis that they need. One of the things that we find, especially in the Microsoft stack from an ETL standpoint, is recently Power Query has a lot of applicability for ETL, and that does help streamline some of the processes, but it also creates an audit trail. That is really significant, that we can create this really valuable audit trail and then go back into that process and either show, “Here’s what we did to go from source document to our tax file,” and then also make changes along the way.

Benefits of New Technologies

Levin-Epstein: Sounds like tax departments can expect some substantial savings by using ETL tools and processes. Is there anything else that provides that kind of an advantage for tax departments who utilize these new technologies?

Panko: I think the advantage for tax departments to use, or at least identify, processes they can apply ETL tools with, the advantage is obviously we’re going to save time. But also we’re going to build out consistency in processes that can translate across movement of people, right? So, our tax departments aren’t always static in terms of people staying there for decades and decades. They’re moving to different jobs and functions and roles. So, by building out these processes and also putting some automation on top of them, we’re building that virtual employee that can handle these processes. And then, the people that are managing and doing the more value-add [work] of analyzing the planning pieces or if they’re compliance correct, that is the value-add service that the tax firms can provide instead of the low value of just doing the ETL process. It’s building up the process so there’s a little bit more automation, and as we get more reporting requirements from all of our revenue authorities, if it’s federal, if it’s state, if it’s international, it gives us the ability to handle that volume and do that on a more consistent basis.

A Quick Nod to the Future 

Levin-Epstein: What would you say is the time frame for more and more tax departments to really adapt to ETL and AI capabilities? 

Panko: I think we have to distinguish between [them] when we’re talking about AI and automation. AI, if we talk about the ChatGPT and OpenID that’s out there, that’s one area that provides maybe just informational pieces. But when we’re talking about automation of our processes, I think that’s going to be something that we can attack now. It’s really kind of depending on the workload of what we have and then we can carve out the time to really do some valuable design and development of it. I think the AI tools that are out there in the marketplace right now are beneficial for looking at background information, but they’re really not going to go through—yet—doing the processes. But there are tools today—and I’m talking more in the Microsoft stack, for example, where we can take OCR [optical character recognition] PDF data, strip off the data in an invoice or a contract, and put that in our analysis, [whether] it’s in Excel or a similar kind of tool.

There’s AI tools already out there in the Microsoft stack that tax departments can use. I think that’s more valuable than kind of trying to figure out how OpenAI or ChatGPT works for us, because those are going to be with our Office 365 tenant; they’re going to be inside our firewalls, so we can actually go and use some of those AI tools that can do the extraction of data and handle the large volumes. We can do those things today. Really, the biggest challenge is finding the time and finding the right people in our organizations to do that for us.

Levin-Epstein: Anything else you want to add?

Panko: I would think, as I said, we’re specialized in the Microsoft stack, so we’re not in other services, just because that’s been our specialty. And most tax departments, especially within TEI, are going to be on a 365 tenant. I think one of the things to encourage individuals is to get access to all the tools. Our internal IT departments don’t always want to give us access to all the tools, but there are some really valuable tools within our 365 cloud that are inexpensive to use and can save us a lot of time and provide more of an audit trail and handle a lot more volume. It’s something that we really have to talk about with our technology teams to give us more access, not necessarily to certain information, but to the better tools, and really getting better access going forward.

Levin-Epstein: Thank you

Leave a Reply

Your email address will not be published. Required fields are marked *

XHTML: You can use these tags <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>