I've been in a lot of conversations about AI lately. Conferences, panels, hallway chats with founders and operators. Everyone's talking about how AI is going to transform supply chain and warehouse operations.

And honestly? I'm still making sense of it.

I'm not an AI expert. But I've spent enough time in warehouses and around automation vendors to know when something feels like real operational value versus just impressive technology. So I'm trying to parse through what I'm hearing using the only lens I have: what actually helps operators ship orders.

Here's what I'm learning. And what I'm still trying to figure out.

Back in October, I moderated a panel at a robotics summit where NVIDIA's GM made a bold prediction: Physical AI will be the next trillion-dollar industry. He painted a future of 2 billion cameras, 200,000 warehouses, and a billion humanoid robots - all powered by AI that moves robots from task-based programming to intent-based understanding.

One demo stood out. Instead of programming a robot to pick a specific object, the instruction was: "Pick up the object on the table that can be used for drinking water." The robot correctly chose the tumbler over a mug and a hairspray can.

That's not small. If robots can understand intent instead of just following programmed tasks, that could dramatically shorten training and deployment curves for warehouse operators.

But then came the grounding moment. The CEO of Plus One Robotics noted that from the moment a truck is unloaded to when a package reaches someone's doorstep, there are still 22 human touches.

Even as the technology accelerates, human judgment and labor remain deeply embedded in every supply chain.

My takeaway: Technology is moving faster than ever. But the real bottleneck isn't compute power - it's human adoption. Warehouse operators aren't paid to deploy robots. They're paid to ship orders.

Around the same time, companies like Augur, Augment, and Pallet started showing up everywhere in my feed. Lots of hype, lots of funding announcements, everyone calling them "AI for supply chain."

But when I try to understand what they actually do, I keep hitting the same question: are these painkillers or vitamins?

Someone posted after Manifest that these tools seem to be layers over traditional enterprise systems. If that's true, then they're enhancing existing workflows, not solving acute operational pain. That's the vitamin vs. painkiller distinction - nice to have versus can't live without.

I don't know yet. I haven't used these tools. But the question matters because it gets at something deeper: who are we building AI for?

Here's where I'm trying to apply my operator lens.

Most of the AI conversation in supply chain is about helping enterprises make decisions faster. Better forecasting, smarter inventory allocation, faster exception resolution at scale.

That's all top-down thinking. C-suite problems. Enterprise optimization.

But what if we flipped it?

What if instead of asking "How can AI help the enterprise make better decisions?" we asked "What does the warehouse manager want AI to take off their plate so they can do more of what they're actually good at?"

Because here's what I keep seeing: operators are drowning in work that doesn't require their judgment. Report generation, data reconciliation, scheduling coordination, chasing down exceptions that could be triaged automatically.

The things that actually need a human - coaching associates, handling the weird edge cases, making judgment calls during chaos - that's what gets squeezed out by all the administrative noise.

If we could use AI to handle the noise, operators could focus on what they're good at. And if individual operators are working at their highest level, the enterprise wins more sustainably than any top-down optimization could deliver.

That's the thesis I'm working through. But I'm not sure I'm right.

Are tools like Augur, Augment, and Pallet actually solving for individual operators? Or are they solving for executives who want better dashboards?

What are the tasks warehouse managers are doing today that AI should take over? Not in theory - in practice. What's the list?

Is the real opportunity in enterprise-level AI or individual-level AI? Or is that a false choice?

And maybe most importantly: how do we avoid the same pattern we've seen with automation, where we celebrate potential instead of performance because we haven't aligned the technology with operational incentives?

I don't have clean answers yet. But I'm paying attention, and I'm trying to figure out where the real operational value is versus where the hype is.

If you're working on AI tools for supply chain, or if you're an operator who's tested these systems, I want to hear from you. What's working? What's not? What do you wish AI would take off your plate?

And if you're heading to MODEX in April, let's connect. I'll be there doing the same thing - asking questions, learning from operators, and trying to separate what's useful from what's just impressive. Atlanta's my backyard, so reach out if you want to compare notes.

We're all learning together on this one.

-Parth

Thoughts? Questions? Feedback? Reply to this email.