Rather watch?

Video transcript

Transcribed from this video. May contain transcription errors.

My name is Caleb Gilbertson. For those of you who don't know me, I co-founded Imprint Engine in twenty twelve. We actually celebrated our fourteenth anniversary last month. Since then, we've grown to more than one hundred and fifty people with teams around the world, the US, Dublin, India, Australia, and the UK, and we're still growing quickly.

Over the last several years, one of the things I've spent a lot of time thinking about is AI. A couple of years ago, we also co-founded a company called MerchAI. The idea was simple: build AI tools specifically for the merchandising industry and see where this could take us. And the deeper we've gotten into it, the more convinced I've become that there's an enormous amount of opportunity ahead. But there's also a lot that all of us in this room are going to have to figure out because the way our industry creates value is about to change.

So with that said, let's get into it.

I remember a world before the Internet. Dial-up tones, AOL disks in the mail, minutes that felt, felt precious. Then cell phones, and then everything connected. And in what felt like a blink, we went from hunting for information to drowning in it.

A year ago, I wrote about swag twenty thirty-five, and some of that future isn't here yet. Uh, but think about what has already shifted. We went from AI that gives answers to AI that gives reasoning to AI that takes action. That acceleration is, is real.

Now, imagine twenty thirty-five. You get in, and a customer sent a simple message the night before. "We're onboarding five thousand people globally. Uh, we want it to feel meaningful. Here's the budget. Show us what you would do."

That's it. By the time you read it, systems have pulled in brand history, prior orders, engagement data, what's in inventory, what's stuck in customs. Uh, you wake up to, to decisions, like not tasks. And that's the arc we're on.

Uh, so hold that twenty thirty-five image for a moment. Um, the question is, yeah, what does that make valuable?

Uh, before we go there, yeah, what will AI never replace? I hear relationships, creativity, judgment, empathy.

Yeah, here's a test. Yeah, whenever someone says, "AI will never do X," ask, "What would AI need in order to do that?" Knowledge, context, uh, reasoning, uh, tools, agency.

Now try relationship, memory, follow-through, responsiveness, trust. Which of those is fundamentally impossible for a machine?

That's where it gets uncomfortable. Um, not because AI is human. It isn't. But because of consequence. Machines can model, but humans bear the stakes. Humans decide what's worth optimizing. Humans make promises.

So, here's the pivot. If intelligence becomes cheap, what becomes valuable? Information gets cheap. Coordination gets cheap. A lot of execution gets cheap. Value moves upstream from doing to choosing.

And for us, that means having the ambition to take responsibility for a bigger customer problem.

Think about that, uh, onboarding example. Say the customer wants welcome kits. If an agent can handle sourcing and spit out a quote, uh, what do we turn that time into? Send the customer twenty options by lunch or ask, "What should someone feel on their first day? What has gone wrong before? What needs to be different across five countries?"

That's where I want us to go. Get involved earlier, understand the purpose, help shape the experience, and then own the delivery and what we learn from it. Concept before product.

So if AI gives your team time back, uh, what are you going to turn that time into?

Uh, that changes the distributor model. Yesterday, we sold products. Today, we manage projects. Tomorrow, we design outcomes.

A few predictions to make this concrete. Quotes in seconds, not days. Product search as we know it disappears. Suppliers will increasingly sell to machines. Revenue and headcount decouple, and context, um, your proprietary memory and judgment, that becomes the asset.

The biggest threat isn't AI, it's defending value that's already gone.

So I come back to you. What will AI never replace? Whatever your answer is, don't ask if AI can do it today. Ask what it would need and what happens when that exists.

Now, the future isn't about AI, it's about the reinvention of value, and it will belong to the ones willing to reinvent it.