AI, ERP, and the Modern Manufacturing CFO: Part 1 of 2 

In part one of a two-part interview series, Jim Gross, Client Account Manager for NexTec’s Acumatica Cloud ERP customers, shares his perspectives on how the convergence of AI and ERP is changing how manufacturing CFOs and finance teams work. Before he began his ERP consulting career, Jim spent nine years working as a Controller and as CFO at a fast-growing mid-sized manufacturer. Here he discusses the fundamental changes that are coming to the manufacturing finance function with rapid advances in AI.  

Jim, take us back to your time in the CFO chair. What were the most common challenges caused by disconnected systems, spreadsheets, and manual processes?

Going back many years, all we really had available to us was daily transaction processing, month-end journal entries, account reconciliations, monthly close activities, financial reporting, and things like that. A common rule of thumb was that we spent 80% of our time looking backward and 20% looking forward. We weren’t predictive. We weren’t looking at future trends. We were just explaining what happened in the previous month or quarter. That’s what we were tied to back then. We didn’t have the tools to look at anything else. 

What’s the real dollar impact you’ve seen when a manufacturer moves from fragmented systems to one single source of truth with an ERP platform? 

It varied, but in my space, we could easily be losing 20 to 25% of the costs or budget we were expecting, because we didn’t know what was happening on a day-to-day basis. We were always going back and trying to find the holes in the transactions that caused us to lose money. Many times, it meant going back to a supplier and finding out they had overcharged us in a previous period. The dollar impact was significant enough that we were literally spending the entire month going back, trying to reconcile the previous month, and then trying to prevent it from happening again. 

Can you walk through a specific example of a mid-market manufacturer who is seeing real results from AI? 

That’s really where a lot has changed, because there’s so much automation in modern ERP systems. Invoice processing, expense management, bank reconciliations: those are all part of the real-time data now.  

But here’s where we’re really seeing the effect. As a manufacturer, we used to capture labor on every production order and then wait until the end of the month to find out, say, that Steve didn’t book his time on a production order on day three. Now we can see that happening every single day. When you look at a production order in the system, you’re not only seeing the six steps it might take to close it out; every step gives supervisors or management an opportunity to dive in and do the reconciliation in real time. If somebody missed clocking in or out on a job, you can go right there and ask, “How much time did you actually spend on it?” “Oh yeah, I clocked in at 8 and forgot to clock out at 10.” You can see those gaps in real time and make the changes. 

So, as that production order moves along, you can see whether you’re hitting the gross margin percentages you planned for, and whether you’re using the raw materials you planned for. You can see scrap factors in real time, you can see how the production order closes, and you’re watching the variances happen in real time across all the steps to complete it. 

How much can AI reduce the time it takes a manufacturer to close its books each month, and catch any inconsistencies before they reach the auditors? 

You’re moving from what happened last month to what’s happening now. Instead of waiting for a month-end report, executives can see performance daily. They have dashboards available to them. They can drill into daily sales. They can dive into daily production orders. The financial data is integrated across all functions. 

Let me paint a scenario: a manufacturer that is also a project-based company. The project manager working in the field with their team can see where a production order is and what state it’s in. Are they waiting on raw materials? Are they waiting on goods to be completed in the factory? They have access to see everything in real time, so they can make better decisions in the field. The manufacturing folks get a better sense of when production orders need to be completed.  

I’ll give you another example: Rather than saying, “We missed revenue by 5%,” AI can help you identify which customers, products, and regions caused that variance on any given production order. 

That’s done using the AI anomaly detection in today’s systems. You set your parameters. Say you want to hit a 35% gross profit on a particular product, or a particular customer and product. You see that happening as the product moves forward in the system. Maybe a supplier you normally rely on accounts for 10% of your costs, and they increase their price to you, which will now throw your gross profit off. You’re catching that in real time, at the purchase order receipt stage, and the system flags that it’s above the variance tolerance you set for raw materials coming in for this project. 

The ERP system is now alerting you based on the business events and notifications you’ve set up. You define the parameters for every sale, every production order, each product, and each customer. When costs go outside those parameters, the system sends you real-time alerts saying, “We have a problem.” You can address it right then, at that moment. You can adjust your production order or go back to your suppliers and see if you can get the price they promised you.

A lot of CFOs still think of ERP as just an accounting system. How do you explain to a skeptical finance leader that ERP now spans operations, production, inventory, shipping, and so on? 

That’s the biggest challenge. When we’re talking to a company coming off a homegrown system, or a legacy system that hasn’t been updated in years, that’s their mindset: “We’re just going to replace the system we have.” 

We try to explain that that’s fine for go-live, day one. I get that you want to be able to enter orders, invoice customers, pay your bills, and get your financial statements. But phase two is taking advantage of all the features in a modern ERP system. 

This is where we have an opportunity as a VAR to educate our clients on the next steps. Here’s a quick checklist. A traditional CFO was a scorekeeper; a modern CFO is a strategic advisor. The traditional CFO looks at historical reporting; the modern CFO looks at predictive insights. Cost control versus value creation. Spreadsheet-driven versus data-driven. Annual planning versus continuous planning. We have an opportunity to educate our clients and make them aware that these tools are inherent in their system. Let’s set them up so you become that more modern CFO. 

With a finance leader who still thinks of ERP as basic bookkeeping, are there “wow” moments? Are there things you show them that are real eye-openers? 

Absolutely. That’s the cool thing. When the system starts emailing them to say, “We have an issue,” or “We have a flag in the system,” because they asked it to identify anomalies, that’s when they gain confidence in the system. When they drill in and see that it’s all data-driven, based on what their own team put into the system, they have more confidence that it really is a modern system that’s going to give them the tools to make adjustments in real time. 

The other thing is this: I’m old school, I’ll tell you right now. Traditionally, you’d close the sub-ledgers in the first 5 to 10 days of the month and then spend the next 10 to 15 days closing the financial books. When you see that the sub-ledgers can be closed on the last day of each month, you’ve just gained 10 to 15 days. Instead of spending 10 to 15 days doing reconciliations, they’ve already been done.

Tariffs and margin pressure are top of mind right now: How is AI-enabled ERP changing the way finance teams track landed costs and respond to supply chain and market volatility in real time? 

Now you can predict those costs. You’re not waiting for an invoice from a carrier that brought items in from another country. You’re not waiting 15 to 20 days after that. You can get quotes from those carriers into the system, and that can be part of your purchase order and your landed costs. A purchasing system lets you add those costs to your inventory cost. 

So, when you’re selling a product, you’re saying, “The actual cost of the material is $10, but it’s going to cost me $5 per unit to get it into the country.” Now you’re looking at that and saying, “I need to increase that cost to $15,” which will change your selling margins and your sales price. It changes all of that. With landed costs in real time, you’re putting that in before the order physically comes in the door. So now you can see that exposure, and all you do is true up the differences when it arrives. 

You’re no longer waiting four or five weeks. Because what happens today is that if you receive the physical inventory, but don’t get the bill from the carrier until four or five weeks later, you’ve probably already sold that inventory, and you can’t go back and recapture that cost. That’s all changed with modern AI-enabled ERP. 

Forecasting used to mean pulling data into spreadsheets and rebuilding assumptions from scratch. How does having AI models connected to live ERP data change those conversations with the C-suite, board, or ownership? 

It’s now self-serve. Inside an ERP system, you can build your data views and your generic inquiries (GIs), have AI help build part of those GIs, and now, with Copilot built into these systems, you can ask the questions you need to ask. 

With some of the modern data tools out there, like Power BI and all the different BI tools, what you’re seeing is more forecasting. It’s more looking ahead, using trend analysis to get you to that forecast. With the system able to show you trends for the different products you sell or the different customers you sell to, you can start to predict what the future will look like. 

Budgeting was horrible, I’ll be honest with you. The worst part of my job as CFO was budgeting, because it was nothing more than throwing darts at a dartboard. No salesperson really had any data to tell me why they thought next year would be a 3% increase or a 5% decrease. It was hard. Now you can use the forecasting tools in an ERP system, which pull from historical data and show you predictive analysis. What does it look like if costs go up 3% next year? What if my suppliers start charging me 2% more? What if sales go down 5%? You can put all those predictive tools together and let the system help you forecast. Budgeting becomes much easier. 

Can CFOs trust AI recommendations inside their ERP system, or do they still need human review? 

When we think of the word “trust,” we might say, “I’m expecting it to give me all the answers.” It can’t do that. We have to understand that it’s going to give predictive analysis, and that there’s a hierarchy of trust. You have your general ledger, financial statements, ERP reports, and analytics, and then you’ve got the AI recommendations. The closer you get to the recommendations, the more judgment is involved. That’s just the reality of it. 

What AI is very reliable at is pattern recognition: identifying things like unusual expenses, detecting inventory anomalies, flagging duplicate payments, and predicting customer payment behavior. AI gives you the data behind any of these predictive analyses, and you can drill back into it and see, “This was a human error on my team’s part,” or “This is a trend now, because we’re moving into higher costs,” whatever the case may be. So, the trust comes from drilling into the data it provides and proving it: proving that it’s something that happened in your system or through human intervention. 

AI is not going to replace anything. It’s no different than a pilot in an airplane: you still need a copilot. AI becomes the copilot, but you’re still the person navigating your financial system and making sure you’re on course to reach your final destination.

How will AI change the role of the manufacturing CFO over the next three to five years? What’s the biggest shift you expect to see in what it means to be a CFO or finance leader at a manufacturing company? 

I really look at it as the continuous close concept. Traditional CFOs know only the monthly close. With a continuous close, you’re literally closing every sales order, production order, labor transaction, and material move on a daily basis. There should be no reason at the end of the month to go back and reconcile 30 days’ worth of data, or something from 29 days ago. The continuous close model streamlines things so you’re seeing everything daily. 

This is where it really becomes important. On day zero of the close, when your month ends, the reconciliations are done, entries are posted, variances have been analyzed, and finance review and certification are done as well. You’re shortening the close cycle: what used to be a 10, 15, or even 20-day close is now one to three days. You should be able to produce your financial statements within the first week of the month. You should be able to confirm your inventory is accurate. With a continuous close process, every sub-ledger in the system is being recognized and reconciled daily. 

When some finance people hear “AI integrated into the ERP,” they worry it’s coming for their job. Will AI replace finance teams, or just change what they do?  

Instead of replacing jobs, it’s going to make people more productive and more useful to the company. The reality is, I would have loved to be able to forecast more easily back in the day. Instead of being a CFO reporting month-old data to the company, I want to be a visionary. I want to be a financial planner and analyst. I want to show the company where we’re going in the next three to six months, based on predictability, anomaly detection, and all the forecasting tools now available. So, I look at it and say, my job is even more secure, because I’m adding value to the company. I’m not just reporting historical data; I’m adding value by looking ahead and helping forecast. 

That can happen in any function in the company, whether you’re an AP supervisor trying to save the company money by taking advantage of discounts, or an AR person managing receivables and collecting money faster. I used to pull up an AR aging report and sit on the phone every single week calling customers who were past due. What a waste of time that was. Now the system can auto-generate a past-due statement, with links directly to that statement so they can pay the bill online. My collection days are gone. I don’t have to worry about that anymore; the system is doing it. That’s where I become more productive. I make better use of my time, I can better save the company money, and I can be more a part of the senior management team. 

Jim Gross is a Client Account Manager in NexTec’s Acumatica ERP consulting practice, working with clients to help them implement and optimize their core financial systems and processes to support sustainable growth.

Jim has been working with Microsoft Dynamics SL since 1989 and Acumatica since 2017. Before he began providing management and financial systems consulting, Jim spent 9 years in corporate management, working as a Controller in the service industry and as CFO at a fast-growing manufacturing firm, where he led the selection and implementation of an ERP system to handle the growth of their business and manage their vast product inventory.

With nearly four decades of experience with financial and ERP systems, he brings a wealth of real-world expertise to every Acumatica implementation. Outside of work, Jim loves golfing, hiking, taking beach vacations, and spending time with his wife Missy and their two kids and four grandkids. He also travels extensively doing mission work in the US and abroad.