Why AI in Supply Chain Needs Real-Time Execution Data to Deliver Results: Q&A with Saumya Saxena
AI is getting a lot of attention in supply chain, but initiatives often fall short of expectations. Across operations, there’s often a disconnect between what’s happening on the floor and what’s captured in the system. Transactions are entered late, updated in batches, or handled through workarounds. That gap makes data harder to trust and limits how useful it is for decision-making.
In this Q&A, RFgen Solution Architect Saumya Saxena breaks down why real-time execution data matters, how gaps in data flow impact AI outcomes, and what organizations need to address before scaling AI across their supply chain.
Q: Why do so many AI initiatives in supply chain struggle to deliver real results?
Saumya:
A lot of it comes down to how AI is being approached. It’s often treated like something you can layer on top of existing operations, when in reality it depends entirely on how those operations are functioning underneath.
In most supply chains today, there’s still a gap between execution and systems. Data gets captured late, sometimes manually, and not always consistently. There’s a delay between what happens physically and what shows up in the system.
AI doesn’t fix that—it brings it to the surface. You’ll still get outputs, but they won’t reflect reality. And that creates a false sense of confidence, which can actually be more dangerous than not having AI at all.
Q: There’s a lot of pressure from leadership to “do something with AI.” What are executives often overlooking?
Saumya:
They’re underestimating how much operational discipline AI really requires. There’s this idea that AI will automatically create visibility or intelligence, but AI can only interpret the information and processes it’s given. If the underlying data isn’t accurate or timely, the interpretation falls apart.
What I’m seeing is that leadership teams want faster insights, but they don’t always have specific goals for AI clearly defined, or don’t always know what foundations are required for a successful implementation of their AI initiatives. Because of that, they’re not always investing in the infrastructure needed to produce reliable data. So, you end up with dashboards and predictions that look impressive, but the teams in the warehouses and in the corporate offices can’t trust that the data represents what’s happening on the floor.
Q: How does poor execution data actually impact AI outcomes?
Saumya:
It tends to snowball. If your inventory data is even slightly off, that impacts fulfillment decisions, which leads to delays or errors. Then teams step in with manual workarounds to fix those issues, and that introduces even more inconsistency.
Now your AI system is learning from that environment. It’s not just dealing with bad data, it’s picking up on unstable patterns. Instead of improving performance, it can start reinforcing inefficiencies.
That’s why data accuracy isn’t just a technical concern. It directly affects cost, customer satisfaction, and how resilient your supply chain really is.
Q: What does “AI-ready” actually look like in a supply chain environment?
Saumya:
At a high level, it comes down to having a continuous connection between execution and your systems. Every transaction, whether it’s receiving, picking, or moving inventory, is captured right when it happens and reflected immediately in the system. There’s no lag, no batching, and no reliance on updates later.
When you have that kind of real-time data flow, AI can work off a live and accurate picture of the business. That’s when it starts to deliver real value, whether that’s better forecasting, faster decisions, or more effective automation.
Q: What’s the most effective way for organizations to move toward AI maturity?
Saumya:
Start by improving how data flows through the organization, rather than jumping straight to new tools. That means expanding real-time data capture across operations, reducing delays in how transactions are recorded, and making sure your systems are tightly integrated so information moves seamlessly.
It also helps to define clear operational metrics and what the organization hopes to achieve with its AI initiatives. This helps companies appropriately phase their approach and have documented measurements in place.
And from there, it’s really about taking an incremental approach. Start with a specific area, prove the value, and then build on that. Trying to tackle everything at once usually just creates more complexity.
Q: How do you see AI reshaping supply chain operations in the years ahead?
Saumya:
AI will definitely play a major role, but I think the bigger shift will be in how organizations think about data. We’re moving toward supply chains that operate more like continuous, real-time systems instead of disconnected processes. AI will sit on top of that and help drive faster, more informed decisions. But the companies that benefit the most won’t necessarily be the ones that adopt AI first. They’ll be the ones that invest in building a strong operational and data foundation underneath it.
Q: What’s one thing leaders should rethink about AI in supply chain?
Saumya:
AI isn’t a shortcut to better operations, it’s a multiplier. If your operations are strong and your data is reliable, AI will amplify that and help you move faster and make better decisions. But if those fundamentals aren’t in place, it will amplify the problems just as quickly.
