Power

Note: If you want to help shape how this agentic version of Intelligems looks like, reply to this email. I'd love to talk with you about it.

When Adam and I started Intelligems, we made a decision that felt counterintuitive at the time: go SMB first.

The gravity in pricing software pulls you toward enterprise — bigger contracts, higher retention, more complexity, the kind of customer who'll pay for a six-month implementation. We chose the opposite. Not because we didn't want those customers eventually, but because we had a theory: if you build something simple enough for an SMB to use on their own, you can always add complexity on top. But start with a complicated, custom thing built for one big customer, and you can't generalize it; you end up rebuilding it for each new customer.

So, we kept it simple. And then our power users started breaking it.

Not breaking it literally, but the agencies and operators who found us early started using Intelligems in ways we hadn't exactly designed it.¹ We realized that having a modular product — one that was "hackable" to different use cases — could be a strength. Put it in the hands of power users and they'll make cool shit.

We knew this intuitively — it is how we wanted to build from the beginning — but maybe it should have been more apparent. Ecommerce is a channel that was built because of technology. And technology always gets pushed forward by power users. The instinct to hack, to build the thing because "it would be so cool," is what evolves the tech and sets the best practices that everyone else follows.

An example: When Glossier transitioned from "just a blog" to one of the hottest DTC brands, they built out a massive internal engineering team to push on editorial- and community-driven commerce. The full vision didn't quite pan out in the way they hoped it would, and, when they migrated to Shopify a couple years ago, a bunch of people pointed to their previous attempts at building custom as a mistake and a sign of naivety.

I think the alternative view is that they were just being an ecommerce brand, and the fact that Shopify built that infrastructure better, cheaper, and at scale doesn't change the underlying truth that technology in ecommerce can be a competitive advantage.

What's changing, though, is where that competitive advantage lives. Shopify (famously, and perhaps pertinently for this newsletter, itself started because Tobi Lütke wanted to sell snowboards on the internet) has largely solved the ecommerce infrastructure problem.

The place to gain advantages is changing.

Back to the Glossier example for a moment: When they migrated to Shopify, they did so with a radical UI design on their homepage that perfectly fit their marketing campaign at the time. I doubt they would have ended up there without having pushed the limits themselves for years.

And this is where the advantages are now: The experience for the customer.

We see this inside our own product, and our power users are showing us where those advantages are emerging.

The tradeoff, though, is real. A product built for power users is a product that can feel overwhelming to everyone else. We've felt that tension: The same composability that lets an agency string together a custom merchandising test is the thing that makes a first-time user wonder where to start. We're improving here, but I don't think the answer is to soften our edges. We don't want to have to choose between "easy to use" and "powerful."

The answer is to hold both: infrastructure deep enough for the builders, surfaces simple enough for everyone else. That's what a real platform does. It doesn't force a choice between depth and accessibility. It lets the complexity live where it belongs — underneath — and meets users wherever they are on top.

The reason this feels urgent right now is that the population of people who can build on solid infrastructure — the number of true power users — is expanding fast. It used to be that "power user" meant someone technical enough to find the edges on their own. Now it's anyone willing to have a conversation with Claude Code or their agent of choice. The barrier to building is dropping, which means the value of having something worth building on is going up.

We're going to keep making Intelligems modular and, increasingly, externalized. We're making our infrastructure available through APIs, agentic surfaces, software that comes to you rather than requiring you to come to it.² We're going to let the power users build their wildest dreams, and then harden what works.

Our bet is that the instinct to tinker doesn't go away. If anything, more people have it now than ever. And we want to be the infrastructure that makes everyone a power user.

1  And this still happens a lot today: customers pull test group assignments through the JavaScript API and pipe that signal into their own custom apps. They build two separate Shopify collections with different sort logic, run them as a template test inside Intelligems, and — suddenly — they're A/B testing merchandising strategies. They construct fake tests just to route data somewhere they actually want to analyze it.

2  We released an MCP and APIs this month. If you're already building on top of Intelligems — or want to — reach out. Hundreds of orgs are already doing it, and we want to know what you're making.