Rebuild

Why we rebuilt the entire agency around going deeper — and why every agency will have to

Most agencies are running a model that’s already dead. They just haven’t noticed yet.

The model goes like this: a company hires an agency to do a thing — run the ads, write the blog, manage the social, build the site. The relationship is topical. It lives at the surface of the business, in a shared drive and a monthly call. The agency knows the brand guidelines and the campaign goals, and not much else. When the contract ends, almost nothing of what the agency learned stays behind, because almost nothing was ever really captured.

That model wasn’t stupid. It was rational. Going deep into a client’s business — understanding the stakeholders, the workflows, the why behind every decision, the institutional knowledge that lives in three people’s heads — costs an enormous amount of time. And time was the one thing an agency couldn’t spend freely, because it had to be spread across a roster. So agencies stayed shallow. Not because they didn’t care. Because depth didn’t pencil out.

That constraint is gone now. And once you really sit with what that means, you can’t un-see it. So we’re saying it out loud and burning the boat behind us.

The bottleneck was never talent. It was context and capacity.

Here is the thing almost everyone is getting wrong about AI in our industry. They’re asking, “How much of this can we hand over to the machine?” As if the goal were to remove people from the work.

That’s the wrong question, and it leads somewhere hollow. If your entire move is to automate the human out of marketing, you end up producing more of the exact thing the world already has too much of — generic, sourceless, forgettable content that could have come from anyone, about anything.

The right question is the opposite one: How do we use this to bring more of the right people, more authentically, to the right place?

We wrote a business plan around that idea before AI was even part of the conversation. The whole premise of HandBuiltBrands was that people will always want to buy from people, and that authenticity isn’t a tactic you bolt on — it’s the only durable advantage there is. Everything we believe about AI now is just the next turn of that same screw. AI doesn’t replace the relationship. It finally makes the deep relationship affordable.

What AI actually removed wasn’t the need for marketers. It removed the operational drag that kept good marketers shallow. It gave genuinely talented people superpowers. And superpowers change the math: when one team can do far more, the smart move isn’t to take on more clients — it’s to go deeper with fewer. To stop being a vendor with a topical relationship and become an embedded partner with a real wedge into the company.

The unglamorous part is the whole game.

If you want to be a true partner — not a vendor — you need an almost embarrassing amount of context about a company. And capturing that context, storing it, and keeping it alive is the least glamorous work in marketing. It’s also the entire moat.

We call it the bedrock and the foundation. Before we build anything for a company, we build a knowledge base about it: who the stakeholders are and what they care about, how decisions actually get made, where the institutional knowledge lives, what’s been tried, what’s been promised, what’s true. We map workflows. We ask a lot of questions — about process, about handoffs, about the parts of the business everyone assumes are obvious. We’re naturally curious, and it turns out curiosity is now a competitive advantage you can actually monetize.

This is also why an “AI assessment” is usually the wrong first ask. A company comes to us wanting to know what they should automate, and the honest answer is almost always: start with the business first — because the fastest way to waste money on AI is to build something impressive that the business never needed in the first place. You can’t automate a process you’ve never written down. The foundation comes before the build, every time.

Being remote from day one turned out to be a quiet head start here. We never had the luxury of knowledge living in a hallway conversation, so we wrote things down. That discipline — documenting instead of remembering — is exactly the muscle this moment rewards.

We didn’t just adopt AI. We rebuilt the company on it.

This isn’t theory for us. We tore up our own operations and put them back together around these ideas, and we did it on ourselves before we did it for anyone else.

We moved onto a native AI project management system and built out a full client registry — context that, frankly, used to live in one person’s head. For every client we now keep both internal and external context documents: who the stakeholders are, what we do for them and how, the preferences, and the unique characteristics of working with that specific company. We tied our email into an AI assistant, and we have an AI chief of staff — we call him Otto — who helps run our internal project management and gets a little sharper every day.

One discipline holds all of it together. We built an AI process we call Mr. Clean that standardizes how every meeting becomes knowledge — preserving the raw transcript, a clean transcript, the notes, and the action items. The standard I’ve held for years, long before any of this was possible, is simple: someone who wasn’t in the meeting should be able to learn everything they’d ever want to know about it. Now that’s not an aspiration. It’s a workflow.

A lot of this work, honestly, is judgment about where the intelligence belongs. Not everything needs an AI brain. A huge amount of what makes an operation reliable is plain deterministic logic — if this, then that; if this, when that. Knowing the difference, and resisting the urge to put a model where a rule belongs, is most of the craft.

Branding is now a system, not a PDF.

The same shift hit our brand work. It used to be enough to say, “Send over your brand guidelines.” It isn’t anymore. With the design systems now coming out of tools like Claude and Google, a brand can’t just be a style PDF sitting in a folder — it has to be a living design system that machines and people can both build against. So our brand identity process and our design-system creation have folded into a whole new layer of branding. You don’t hand off guidelines anymore. You bake the system in.

We’re a marketing company first — so the marketing is built for this world.

Underneath the philosophy, we’re still natively a digital marketing company, and we’ve been building marketing and sales capability designed for the AI era rather than retrofitted to it.

Some of it is live. We run an SEO and SEO-content system, and a website content program built by AI, for AI — content engineered to be found and cited by answer engines, not just ranked by search engines. One of our first builds was an AEO/GEO tracker that measures share of voice across generative engines, so we can actually see whether a brand is showing up where buyers now look.

Some of it we’re standing up now — the knowledge bases are written, and what’s left is the engineering time to test, iterate, and ship. That includes Addison, an agent for Google Ads audits, account restructure, and recurring optimization. And it includes the piece I’m most excited about: a digital marketing and sales analytics data layer. In a way it feels like we’re building our own analytics platform, but really we’re just getting all the data into one place so you can find correlation, causation, and the true KPIs that actually lead to sales and growth — not the vanity ones. From that data layer comes a media-mix modeler that lets you toggle and automate spend and optimize by platform and campaign, plus custom lead-source attribution and lead-qualification systems that bolt onto a CRM like Salesforce or HubSpot. And on top of the whole knowledge base: live chat agents trained on everything that comes with being a real partner to a company — going all the way back to the beginning of the relationship.

None of these are the point, though. They’re the proof. They’re what becomes possible once the foundation is real.

This is the most human marketing has ever gotten to be.

Here’s where it lands for me. AI didn’t make marketing less human. Done right, it makes it more human than it has ever been able to be — because it clears away the operational weight that used to keep us at the surface, and lets us be genuinely present in our clients’ businesses.

It’s not “how much can you hand to AI.” It never was. It’s “how much more of the right people can you authentically bring to the right place.”

That’s the bet. Fewer clients, gone deeper. Real partnership instead of topical service. A brand that’s built by hand, with more care, because the machine finally took the busywork.

We’ve burned the ships. We’re not going back. And we think, before long, no one else will be able to either.