These days, it’s hard to find the letters “ERP” without the vowels “AI” in their immediate vicinity. You can’t blame enterprise resource planning vendors for focusing on their artificial intelligence offerings – they’re shiny and new, especially when it comes to the generative AI (GenAI) copilots and chatbots that make for such enticing demos.
This is not to dismiss the value that AI stands to ultimately deliver through ERPs and ERP-integrated systems in the supply chain and beyond. In logistics, AI and machine learning are already widely used in route optimization. AI-based analytics can assess supply chain risks and propose alternatives based on everything from delivery history to weather patterns. Among other feats, AI can also generate standardized product descriptions, vet and recommend suppliers, provide summaries of negotiations, and automate processing of goods receipts and delivery notes.
In ERP warehouse-management modules, AI can help optimize inventory, storage, and the pace of replenishment and production. But for that to work, you need reliable, real-time warehouse and supply chain data, ideally integrated with ERP-based demand forecasting and financial systems. Which takes us to a harsh reality that so many companies considering an ERP implementation or major upgrade overlook: useful AI depends on good data.
Explore related questions
Useful ERPs also depend on good data. So often, companies embarking on their ERP journeys don’t have it – or, if the data’s there, it’s siloed and formatted at the whim of various legacy systems. So, the focus should first be on data, then on the ERP’s core and key peripheral systems such as supply chain management, and only then on AI.
Get your data house in order
Let’s say you want to improve your company’s inventory forecasting. If your inventory data resides in half a dozen systems, neither human nor machine intelligence will be able to make much use of it. One of the great advantages of today’s ERP systems are their cloud-based, centralized data models built around decades of industry best practices and data-governance approaches.
Think of ERPs as organizing principles around data strategies that, ideally, unify business processes using technology. But ERP implementations still beg difficult questions: Are there processes and accompanying data structures that provide competitive differentiation and therefore should be preserved? What are your target improvements, where does the associated data live, and how is it structured?
Having AI capable of recommending suppliers and thereby easing the burden on procurement professionals is a good thing. But having your supplier-related data centralized and structured in a way that an ERP’s off-the-shelf or lightly customized dashboard can help a procurement professional make such calls is a vital first step. With the right data available and an ERP-integrated supply chain management solution in place, human intelligence boosts productivity in the near term, and the foundation is in place for AI integration when the time is right.
Remember also that AI systems, even when embedded in ERPs, amount to implementations of their own. They typically require training on proprietary data sets, parameter-based customization, and thorough testing to ensure security and reliability – i.e., that they’re not hallucinating fanciful answers to mission-critical questions. Just as importantly, AI impacts your people.
Don’t gloss over AI’s workforce implications
The data and process rationalizations that come with ERPs represent huge changes for a workforce, and change is hard. Adding AI to the initial mix can compound the issue. Despite many of today’s AI applications being geared toward providing distilled information for people to use in making better decisions or automating tedious work so people can focus on higher-value tasks, people are worried that AI is coming for their jobs. Focusing on the ERP first and only later introducing AI in digestible, strategic servings is, in my experience, the wiser path.
When you do roll out AI features, it’s vital to have a well-thought-out change management plan to guide employees through their evolving roles as well as a communication plan to anticipate questions and concerns and address them as they emerge.
It’s entirely possible – perhaps inevitable, even – that AI integration will one day make or break the ROI of an ERP implementation. But for now, the strategic imperative should be to gain the efficiencies, operational visibility, and process rationalization that ERPs have proven they can deliver. Doing an ERP right means getting the data right. That will open the doors to AI applications as they prove themselves as strategically relevant down the road.
