Uses of AI in financial control explained at Groundbreak.
December 16, 2024
At Procore’s Groundbreak event in Denver, Colo., Joe DiPinto of Premier Construction Software in Markham, Ont., shared his insights into how AI might help contractors with new capabilities for manipulating their finanical and accounting data.
His first point was on cost control, which he boiled down to managing budgets, managing commitments, managing actual costs and managing forecasts. He pointed to forecasting as a potential strong area for AI because it can connect the costs being experienced in operations and the field to the financial planner in real time. With the latest tools, the office can get a sense of whether a project is going to ultimately end up making money or not based on what happened that day. According to DiPinto, this takes a lot of risk out of the construction business by eliminating a common mistake many contractors make. “I’ve seen businesses go bankrupt, literally, because they are not accounting for the unknowns, and that surprises them,” he says. “It leads them to running into cash flow issues, not billing enough and that becomes the end result.” One of the main barriers to this kind of data sharing between the field and the front office is standardization of how each group collects and reports information, and that’s where AI-powered data tools can come in.
“We all tend to struggle with communication from time to time, right?” DiPinto asks. “I don’t think it’s because we don’t want to communicate. No, by no means. We’re just very busy people. We only have a finite amount of time to do the task. We need to in a bit. And if you’re doing complex analysis, or you’re putting out fires on the job site, you don’t have time to chase people down.” DiPinto sees great benefits from, for instance, project managers being able to approve bills and invoices coming in from the work teams right away. “That’s how you get good cost controls that help you forecast.”
DiPinto says AI helps by improving business functions in five different areas: data entry, automation, information document search, reporting and analytics, predictive intelligence and actionable items.
Data entry automation is becoming fairly common. Think of apps that look at receipts and are able to extract the figures, taxes and company names and automatically fill out the fields in an expense form. The technology is a form of AI called optical character recognition, but DiPinto says it can go deeper than what it’s used for today. “Let’s say the vendor sends you an invoice and the line on the invoice says one unit times 500 equals 1,000. We know that’s not right, but if you’re using tools that aren’t intelligent and checking that and it’s just extracting it, who now has to do that validation? You do.” The latest tools will catch that kind of error and flag it for review rather than blindly copying them into the account.
“What about validating vendor statements?” DiPinto asks. “If you purchase a lot of materials, you’re probably getting pages of statements at the end of the month that the finance team is probably having to double check. AI can check if the invoices are in your system or not, and then send them an email saying, ‘Hey, we need these invoices. Send them over.'” The net effect is to save time that people can use for higher value-added work.
One of the most powerful uses of AI in a construction company is the one familiar to many from using ChatGPT: natural language search. Rather than scanning through folders and spreadsheets themselves, project managers can simply as the system questions. What’s the value of that contract? Has the vendor paid against that contract? Show me the specific scope of work on that contract? An AI-powered system will spit out the answer, summarize it, pull the documents for it and say this is where we got it from, DiPinto says. “So again, think about that. What does that look like today? Gotta go through my document folders. Gotta search for the document, gotta read through the document, or maybe upload it to ChatGPT first and let it work for you. But this is tangible stuff that you can be using to cut through the noise. Just save relentless time spent searching for documents and information.
“It’s like your own virtual assistant.”
DiPinto explains predictive intelligence, an emerging power of AI. “It is actually assessing the current state of what’s happened, then forecasting potential problems and potential results for it. Do we have any projects where we’re we’ve gone over budget every month? So I’m going to forecast that trend going forward. Trend analysis is an absolute massive component look through as it can conglomerate and accumulate different data points and assess them forward.”
