By Gavin Verreyne
SVP Professional Services
SYSPRO Americas
In November 2022, OpenAI released ChatGPT to the public, which sparked a firestorm of interest in generative AI unlike anything the technology industry has seen in many years, if ever. It propelled Nvidia — which makes the high-powered GPU chips that generative AI requires to run — to an incredible $6.7 billion in net income in a single quarter that ended July 2023, a whopping 422 percent increase over 2022. Chief information officers (CIOs) of companies large and small are constantly peppered with questions from their boards and the C-suite: How are you incorporating generative AI into the business?
AI and generative AI, in particular, looks as if it will eclipse all three. It’s no exaggeration to say that it will utterly transform how users interact with technology, and that includes enterprise-resource-planning (ERP) systems. Here’s what I see coming in 2024, and perhaps a bit later, for how AI will change ERP.
How ERP is already employing AI
First, let’s talk about what we’re already seeing regarding AI and ERP. Cognitive AI is already enabling ERP platforms to understand and consume data from documents such as invoices and purchase orders. With this ability, ERP can automate processing these documents into transactions, which is not only far faster than manual methods, but also much more accurate.
Predictive AI is also already having a huge impact, due its ability to use historical data to understand trends and make accurate predictions about future events. For example, predictive AI can analyze supplier behavior to estimate when orders will be fulfilled and which suppliers are most likely to miss delivery dates and by how much.
It can also analyze inventory data to predict and optimize when a manufacturer should place new orders and at what quantities. Using quality data and throughput information, predictive AI can even alert maintenance crews when to maintain specific machines to prevent failure, avoiding downtime and expensive repairs.
These capabilities, which many ERP platforms already incorporate, enable manufacturers to increase efficiency, and identify potential negative business issues before they become crises, as well as provide insights into optimizing operations.
They’re valuable, no question. But generative AI promises much, much more.
Generative AI and ERP
Generative AI — such as ChatGPT, Microsoft Copilot and others — is a specific type of artificial intelligence that has the ability to learn and understand complex and subtle patterns in large datasets of text, images, video, audio, even computer code to create new data models with similar characteristics.
Now, certainly there are well-publicized issues with “hallucination,” where generative AI creates references, facts and events that don’t exist, but with the right guardrails, generative AI is enterprise-ready and manufacturers will start seeing it as a necessary component for ERP.
One of the most important capabilities that businesses can expect from generative AI is an intelligent and much more intuitive way of getting the information needed to drive faster decision making.
Instead of building dashboards or clicking through menus, an ERP user could simply type or even verbalize requests using natural language, “Tell me which raw materials in our inventory are the most at risk of shortages in the next three months.” You could even add to that request by saying, “Put that information into a chart, and send me an updated report via email on the first of each month.” By providing context, AI will better serve the actual requests and provide even more accurate responses.
The implications for employee productivity are staggering. Research suggests that, by partnering employees with AI, businesses can expect a staggering 50 percent increase in productivity.
Apply this capability across the board to any task or process that requires collecting and formatting data into a report, and the efficiencies add up to a significant percentage of employee time.
Another application 2024 may bring to ERP, thanks to generative AI, is the ability to summarize, index and organize large amounts of data. For example, if I had 10 gigabytes of unstructured data that I needed to index and summarize, that’s a $200,000- $300,000 consulting engagement. Generative AI could accomplish this task in a small fraction of the time it would take a consulting team to complete.
Within ERP, generative AI could index and summarize extremely complex purchase orders, drawing attention to the most important information for the user, depending on the query. Once we get to expose AI to data sources outside of the organization, the opportunities become limitless.
In fact, generative AI integrated with ERP could take summarized purchasing order information, feed it into a predictive algorithm and then produce a report that forecasts demand for specific products. Considering buying patterns from data outside of the organization, it could also compare the PO data it extracted with inventory and supply chain information in the ERP, and then predict precisely which of the company’s key raw materials is most likely to experience shortages in the next three months.
A human-like user interface
Certainly, users will still want to use menus and look directly at data … for a while. But working with the ERP will increasingly feel like working with a very human-like AI chatbot that could undoubtedly pass the famous Turing Test, an iconic experiment developed by Alan Turing in the mid-20th century to determine when a computer program displays human-like thinking. Dashboards will be simple to set up because users will be able to describe and request changes to them in natural language.
Even better, generative AI in concert with predictive and cognitive AI could suggest additional or alternative key performance indicators (KPIs) to monitor, depending on goals. In fact, a user could ask, “Which KPIs should I monitor to best understand the current level of risk in our supply chain?”
The ERP is the manufacturer’s system of record, and the platform is used not just by people in office buildings, but also by workers in the factory and a multi-generational hybrid workforce. AI and, specifically, generative AI has the potential to make the ERP user interface much more friendly and usable regardless of location or function.
Beware the hype cycle
However, it’s important to remember that we are currently early on the rising edge of the generative AI hype cycle. It’s a feeding frenzy, with every vendor and every enterprise clamoring to demonstrate that it has a generative AI strategy. This means there’s going to be — and, frankly, already is — a lot of AI-washing. Everyone wants to get into the game quickly, which means that some vendors, including ERP vendors, will take shortcuts and market their platforms as powered by generative AI, when, in fact, they fall far short of that mark. Manufacturers will need to evaluate ERP vendors’ AI capabilities carefully. Over time, as generative AI matures, it will become easier to separate the wheat from the chaff in ERP.
Generative AI will transform every interaction we have with software and technology generally. Those organizations that take advantage of it wisely will gain a competitive advantage over those that don’t, but rushing into deploying AI just for the sake of implementing an “AI strategy” or to sell an “AI-enabled” product will do very little to advance manufacturers’ organizations.
Only if AI is employed to solve real business problems will it truly add value. I’m confident that AI and generative AI have the potential to do exactly this, but manufacturers will need to sort through the hype to identify those ERP platforms that have implemented AI intelligently and not just as a marketing gimmick. Don’t underestimate the answer to “Why?” when it comes to AI.
