An interview with Kerrie Jordan, CMO and SVP Product, Epicor
Most distributors have more data than they know what to do with. The problem isn’t access to information — it’s the gap between seeing a problem and doing something about it. That’s the starting point for Kerrie Jordan, who holds the unusual dual role of CMO and SVP Product at Epicor. She leads the company’s AI strategy across manufacturing, distribution, retail, and building supply, and co-authored Epicor’s 2025 Supply Chain Agility Index, a survey of nearly 1,000 supply chain professionals. In this conversation with Ian Heller of Distribution Strategy Group, Jordan breaks down where AI is actually delivering results, what distributors should realistically expect, and how to get started without overhauling everything at once.
The Insight-to-Action Gap
Q: You’ve written about ‘sparking action’ as the real promise of AI in supply chains. What does the problem actually look like in practice?
Kerrie Jordan:
The symptoms are easy to spot. Teams spend more time interpreting data than acting on it. In the 2025 Supply Chain Agility Index — which surveyed nearly 1,000 supply chain organizations — we found that even companies that consider themselves highly AI-ready are still struggling. Forty-one percent cited integration barriers and 40% said a lack of expertise was slowing their decision-making. They technically know what’s wrong. They just don’t know how to act on it fast enough.
Q: Is this a technology problem, a process problem, or a people problem?
Kerrie Jordan:
It’s all three, and that’s what makes it tricky. But it starts with people and process once you have the basic technology in place. Distributors often sit on years of ERP data, and a lot of that data lacks clear ownership. People don’t trust it. And when you don’t trust the data, you hesitate to act on it. AI adoption works best when organizations invest in workforce readiness and change management alongside the technology — not as a follow-on, but at the same time.
Q: What does ‘acting fast enough’ actually mean for a distributor? What’s the real cost of a day’s delay?
Kerrie Jordan:
It means catching problems before they hit your customer or your margin. Think about rerouting orders before a stockout happens. Addressing backorders before the shipment cuts off. Adjusting pricing before margin erosion sets in. The goal is to get earlier into the workflow — not looking back at last week’s dashboard, but seeing what’s developing now so you can get ahead of it.
What the Data Actually Shows
Q: The Agility Index found that more than 75% of organizations rated themselves high or very high on AI readiness. Is that realistic?
Kerrie Jordan:
It surprised me too. But when you dig into it, it makes sense. These industries have dealt with years of supply chain disruption — trade disputes, labor shortages, volatile demand. That kind of pressure forces you to build better data and decision-making infrastructure. The organizations rating themselves most ready aren’t just investing in AI technology. They’re hiring AI-specific talent, building training programs, and laying the workforce foundation that lets them actually use it when conditions shift.
Q: But AI readiness doesn’t equal execution. Where’s the disconnect?
Kerrie Jordan:
That’s exactly right. Readiness doesn’t equal execution. The gap shows up in integration and expertise. You can have great data and still not have the connective tissue between systems that lets AI do something useful with it. And even when the technology works, the organizational habits — the tendency to run reports and review what already happened — those don’t change overnight.
Q: Distribution scored slightly lower on AI readiness than manufacturing or retail in your survey. Why?
Kerrie Jordan:
Distributors operate in a more fragmented data environment by nature. You’ve got supplier systems, customer EDI data, internal ERP, and often other applications that don’t speak the same language. Data flows in and out in different formats, sometimes through email. That complexity slows AI adoption. But it also means the upside is bigger. If you can get this right, the competitive differentiation is real.
Understanding the Types of AI
Q: Generative AI, predictive AI, agentic AI, AI-powered ERP — these terms get used interchangeably. How do you draw the distinctions?
Kerrie Jordan:
Let’s start with what each one actually does. Generative AI creates content — it can build a schedule, summarize emails, generate ideas. Predictive AI uses machine learning to run scenarios on your data. It can forecast demand or identify growth signals. Agentic AI takes it a step further. It can take human-approved autonomous actions within a workflow. It builds on the other types — gathering insights, making recommendations, and then actually doing something about it when you approve. For a distribution executive making investment decisions, that last category is where the insight-to-action gap closes.
Q: So what should a distributor actually expect when they hear ‘agentic AI in your ERP’?
Kerrie Jordan:
They should expect to be able to log in and interact with their system in plain language. No more navigating to specific screens or knowing which report to run. Over time, the system surfaces insights on its own and suggests what to do. You review it, you approve it, and if you want it to handle routine decisions automatically, you can set it up that way. It’s acting as an extension of your team.
Q: And how does Epicor’s predictive product — Grow AI — differ from Prism, the agentic layer?
Kerrie Jordan:
Grow AI looks ahead. It forecasts. It brings in historical and real-time data, runs predictive models, and can pull from outside the ERP as well. It guides strategy. Prism operates in the moment. It’s the layer that acts on those insights — surfacing at-risk orders, pulling the associated customer records, automating multi-step workflows. Grow tells you what might happen. Prism helps you do something about it.
Real-World Outcomes
Q: What are the most compelling customer examples you can point to?
Kerrie Jordan:
Olympus Group used Prism to increase responsiveness to customers, with on-time delivery as the key metric. A typical eight-week delivery window came down to four weeks, and they were tracking a possible 20% gain in operations. They also used the Enterprise Content Management agent to pull specific information from customer contracts on demand — instead of manually searching through document repositories.
Cascade Engineering Technologies used the Prism Knowledge Agent and found it saved their support team members up to 55 minutes per question when assisting employees. That kind of gain matters a lot when you’re onboarding new people in a tight labor market.
And then there’s Madsen’s Custom Cabinets — a smaller family business doing architectural millwork. They were getting spec sheets with three or four hundred cabinet doors, each with different hardware, requiring line-by-line cost estimates. AI could work through all of that spec data and summarize the costs per door much faster than any individual estimator. They were saving hours per person per job. More accurate, faster delivery to the customer.
Q: A CFO pushback I hear often: ‘AI saves time, but that time just gets absorbed into other tasks. Show me the ROI.’
Kerrie Jordan:
That’s a legitimate question. Saving 50 minutes per query is only valuable if it translates into faster order cycles, fewer delays, or better margins. That’s why we focus on embedding AI into workflows rather than deploying it as a standalone tool. When it’s a standalone tool, your teams have to learn a separate interface. You also run the risk of hallucinations because the AI isn’t grounded in your actual ERP data. When AI is embedded, the accuracy improves, adoption increases, and you can actually tie the time savings back to business outcomes.
Tariffs, Trade Disruption, and Speed
Q: Tariff conditions can change overnight right now. What does a company using agentic AI do differently than one that isn’t?
Kerrie Jordan:
When a tariff change happens, someone has to figure out which SKUs are affected, which suppliers, which customer orders. Manually, that takes hours or days. With agentic AI embedded in the ERP, the system can immediately surface impacted items across all of that data. Your team can start deciding what to do — adjust sourcing, change pricing, communicate with customers — right away rather than spending the first several hours just finding the problem.
You can also connect to applications like Epicor’s financial planning tool to model the margin and pricing impact, or the inventory planning application to think through stockpiling decisions and the cashflow pressure that creates. And you can do all of this from within the Profit 21 screen, without bouncing between systems.
The Human Side of AI Adoption
Q: The 2025 Agility Index found that hiring AI specialists within the first 90 days improves implementation speed by 40%. But how do you handle the workers who are worried about being replaced?
Kerrie Jordan:
Start by building trust from day one. Involve employees in selecting the tool. Be clear about the intended outcome. Tie it directly to who you are as a company — not as a separate initiative, but as an extension of your mission. When AI is embedded in the tools your people already use and trust, they don’t have to overcome a learning curve. They’re enhancing what they already know how to do, not starting over.
Find your internal champions early — the people who are naturally curious about new technology. Bring them in before the broad rollout. Let them work with it, shape how the team uses it, and then lead the adoption. That peer influence is a lot more effective than a top-down mandate.
Q: Is AI creating new hiring demand or eliminating jobs?
Kerrie Jordan:
It’s shifting where the talent is. The high-readiness organizations in our survey — 90% of them — are investing in AI-specific roles: logistics optimization specialists, automation engineers. It’s not confined to IT. Procurement, finance, operations — everyone is being asked to understand and work with AI. The skills that matter are changing, not disappearing.
Where to Start
Q: For a distributor still in exploration mode, what’s the most important thing to do in the next 90 days?
Kerrie Jordan:
Start with a workforce and digital audit. Check your data maturity. Understand where your integration gaps are. Ask whether you have the right roles to actually operationalize AI. Look at what your current technology vendors already offer — because the AI that’s embedded in the platforms your teams already know is where you’re going to see the fastest adoption and the most reliable results.
You don’t have to have perfect data to start. The right tools can help clean and standardize data as you go. But you do need to understand your baseline. Once you know where you are, you can reach out to your technology partners, ask what’s available, and start with a trial. The goal in 90 days isn’t to deploy AI across the business — it’s to know where you stand so you can make a smart first move.
Q: Is there a minimum size or data maturity threshold below which agentic AI doesn’t make economic sense?
Kerrie Jordan:
I don’t think so. The Madsen’s example shows that even a small family business can find significant value when they hone in on a specific problem causing a lot of manual work. Size isn’t the barrier. Clarity of purpose is. Find the problem that’s costing your team the most time and start there.
Share this article:
