It's your fault your ERP system isn't ready for an AI Implementation
What is AI implementation?
AI won’t break your ERP. But your assumptions about how AI and ERP should work together might. AI won’t magically clean your data, harmonize your processes, or tell you what to do. It’s not going to turn your 20-year-old ERP into a futuristic command center overnight. And it sure isn’t going to overcome decades of organizational bad habits.
AI implementation is a strategic transformation, not a software plug-in. It demands alignment across people, processes, and platforms—not just tech upgrades. If you’re trying to bolt AI implementation onto a tangled web of disconnected systems, organizational silos, tribal knowledge, and manual processes then you’re not building the future—you’re paving a faster path to dysfunction. AI isn’t failing you, you’re failing AI because you’re not giving it what it needs to thrive.
Let’s talk about why that’s happening—and what to do instead.
1. You Think You Can Plug AI Into Chaos
Many companies want to “try AI” without doing the dirty work first. Clean data? Nope. Defined processes? Maybe. Clear ownership? Definitely not.
AI doesn’t fix complexity—it amplifies it. It will take your bad data and draw bad conclusions from it. It will take your bad strategy and apply it really effectively towards bad objectives.
Your ERP isn’t just a data hub. It’s the circulatory system of your enterprise. And if that system is clogged, outdated, or fragmented across best-of-breed Frankenstacks with no cohesive strategy, AI will expose the mess faster than any audit. Which might be a good job to start AI off with – identifying the mess – but you cannot assume AI will automatically fix it.
You can’t pour jet fuel into a rusted-out pickup and expect a race car. You’ll blow compression and end up with a disabled engine. The foundation matters.
2. AI and ERP Speak Different Languages
ERP systems are designed for stability, compliance, and control. AI thrives on flexibility, experimentation, and pattern recognition. The two can be complementary, but they cannot replace one another. They need to have specifically different roles.
So when ERP vendors say “AI is built in,” you need to ask: “Built into what?”
Often, you’re buying yesterday’s BI with tomorrow’s label. This is a challenge because you’re not actually getting AI or ERP and when you’re looking to get both and integrating neither we have an issue in communication. Business Intelligence is important, and it should serve as a functional component of your ERP that can be enabled by AI. This complex relationship about communication between systems that are dependent on the flow of data.
AI needs context, relevance, and connectedness. It needs clean, cross-functional data that speaks a consistent language. That means possibly investing in:
Master data management (MDM)
Data lake architecture
APIs and orchestration layers
Clear taxonomies across systems
Getting one or more of these pieces in place effectively helps data flow from your ERP into a place that an AI can use it to draw meaningful, accurate conclusions. Until then, your ERP and AI are just awkward roommates—living together but not speaking.
3. The ROI You’re Looking for Doesn’t Exist Yet
Everyone wants to talk about ROI, but few actually define what success looks like.
AI-driven ERP initiatives shouldn’t start with dashboards—they should start with strategic objectives. What do you need to predict, optimize, or automate? Why? And what’s your threshold for value?
If you can’t answer that before buying AI, don’t expect to find ROI after. AI is a tool that needs a purpose, it needs to be pointed towards a problem that it can help solve. Without knowing the issue and the direction, AI becomes misdirected.
AI is also designed to provide iterative and regular insights. Forget 5-year ROI forecasts. That’s legacy thinking. AI demands speed to value—measurable wins in 90-day cycles. That means pilots, not platforms. Phased gains, not big bang transformations. Bite-sized deployments that can be built upon rather than monuments that are unveiled in a premium showcase.
Look for an ROI that can be small and achievable, then build AI to achieve it. Use your ERP to track progress towards it. That’s how you grow an ROI with AI and ERP together.
4. “Jim” Will Wreck Your Strategy—Unless You Include Him
Every factory floor, warehouse, or finance team has a Jim.
Jim’s been doing the same job for 25 years. He knows the process better than anyone. He doesn’t care about your AI implementation—unless it gets in his way. Then he’ll kill your rollout with a grunt and a well-placed “see that’s how it actually works” demonstration of fact and reality.
Here's the twist: Jim is your greatest asset.
He knows what works. He knows what’s broken before it breaks. If your AI initiative doesn’t win over Jim, you’ve already failed because he will become a resistor, and his leadership will become a detractor to your adoption of AI. AI strategy must include boots-on-the-ground stakeholders and influencers who see its value as a tool. Treat AI like a co-pilot—not an overlord.
The next steps are to integrate this knowledge with the data from your ERP so you have a clear model that Jim can see, mold, and modify to his heart’s content. That’s where connecting your ERP to an AI strategy starts to yield real gains in production that increase throughput and decrease waste.
5. Agentic AI Isn’t the Future. It’s Already Here.
Chatbots aren’t cute or annoying anymore—they’re strategic. Agentic AI, like intelligent chat agents embedded in ERPs or messaging platforms, are already creating ROI in Sales, Procurement, and Customer Service. It can provide value more quickly than a Google search, with more informed results and better outcomes.
For example, a good use case for deploying a GenAI agent in Customer Service could be:
Set up a GenAI assistant for support tickets or order status.
Deploy a conversational AI bot that suggests next-best actions based on real-time data.
Use a pilot agent to triage alerts in production or flag inventory issues.
In this scenario, you gain a low-risk, high-feedback environment as a test bed that can prove out the results, and you can deploy it now. Your model can pull customer history, order status, and lead times from your ERP and put them in the hands of your AI, showing real predictive analytics that help resolve customer issues. Your people still need to interpret the AI agent’s results and outputs, that shouldn’t change. But what it can do is provide them with scenarios they can see the effectiveness of on a case-by-case basis. Companies are doing this today.
6. If You Don’t Own Your AI Implementation, You’re Just Renting Hype
AI IP is becoming a competitive differentiator.
The companies winning today are building their own models—or at least controlling their own AI roadmaps. Outsourcing AI to vendors who will turn around and sell the same solution to your competitors only creates a temporary step ahead that can become a step behind in a hurry. This is an even bigger issue with ERP platforms that are all SaaS-based these days – multi-tenant platforms share a space and core code that you don’t want being in the hands of your competitors when it comes to how AI is your competitive advantage.
Your AI strategy should be using the data that comes from your ERP in a way you can isolate from the multi-tenant platform. But owning your AI doesn’t mean building everything from scratch, it means being responsible for where access to it comes from. To do this, start by:
Defining your own requirements
Controlling your training data
Creating your own use case backlog
Keeping your learnings proprietary
You can create an AI model that takes advantage of the information your ERP makes available and builds on a minimum viable product, keeping the cost of deployments low and the value attained high in comparison. Start with one business unit. Prove it works. Fund the next one with the savings. That’s how you build internal momentum without blowing your budget.
7. Don’t Just Integrate AI—Architect for It
You don’t bolt AI onto an ERP like a spoiler on a Civic hoping it’ll make you faster.
You have to re-architect how decisions are made. You have to rethink how workflows operate. You have to move from automation to augmentation.
Your ERP is the system that runs your business processes in a software environment. Your AI needs to be designed in a way that allows for informing the people who facilitate those processes to do so with more information and more options that are aligned towards your strategic objectives. In this one purpose – to enable your people to achieve your strategic objectives for improvement using AI – is the definition of the complexity of the confluences between these workstreams. Don’t underestimate the dependencies among these functions.
This means thinking in terms of:
Feedback loops, not one-off insights
Continuous improvement, not static dashboards
Human + machine teams, not human vs. machine fears
And yes, this includes building an AI governance model to ensure oversight, transparency, and lifecycle management—because your AI today will be outdated tomorrow if you don’t create ways of allowing it to continue learning and adding value. Governance allows you to direct your AI towards the goals you want it to help you achieve in a manner that keeps it under your guidance and control. It’s not about preventing AI from taking over the world, it’s about keeping it pointed in the direction that adds value for you instead of allowing it to provide insights from every direction. Design it for that focus and it will quality results instead of just creating outputs.
Final Take: Stop Waiting for the “Perfect” AI-ERP Solution
Your AI-ERP hybrid won’t be built overnight. It’ll be built through iteration, feedback, and people.
Start small. Don’t be afraid to make mistakes. Deploy a minimum viable product, test it, collect results, re-assess, then improve. Use your ERP to provide your AI with information that can be used to draw more meaningful conclusions then act on them. Identify an area that you want to see improvement that you believe is ready for the gains and give it a shot.
But whatever you do, don’t sit on your hands waiting for your ERP vendor to drop a miracle module next quarter. The companies that are winning? They’re building. They’re testing. They’re owning.
So what are you doing?
Let’s Talk Strategy That Delivers.
Leave a comment below, or if you’d like to discuss how AI can fit into your organization’s ERP strategy, contact us at info@theconfluencial.com.
For more thought leadership on AI, digital transformation, and business strategy, follow us on LinkedIn and YouTube. We post weekly insights to help you navigate the evolving world of Enterprise AI.
