How Zapier Built an AI Marketing Brain

Zapier had a problem hiding in plain sight:

Zapier is a company that helps businesses connect their apps and automate work...

But it was watching its own marketing story get lost every time it moved between teams.

An initial product brief might begin with a sharp strategy and a clear story. Then it would move through product marketing, sales, executives, paid media, social, and events. Each team adapted it for its own work. By the time the story reached the market, it could be a fraction of what the team had started with.

One teammate gave Zapier Chief Marketing and AI Transformation Officer Dan Slagen an analogy that captured the problem perfectly: It was like copying an old video over and over again.

"If you have an old video and you start making copies of it, by the ninth copy, it basically becomes completely blurry," Slagen says.

Slagen's team called it "generational loss." Every handoff created another chance for the original strategy, product truth, and brand story to blur.

And the faster Zapier moved, the more dangerous that loss became. In the old software-as-a-service cycle, Slagen says, sales representatives might hit their stride two months after a product launch. Now, Zapier could already be discussing what came next. Context was changing faster than it could move through briefs and handoffs.

Slagen wanted the opposite of generational loss. He wanted "lossless marketing": a way for the original meaning to survive as work moved across teams, tools, and channels.

So Zapier began building a central AI marketing brain designed to give its people and AI agents the same current context.

The Brain Gives Everyone the Same Starting Point

Zapier has launched an internal V1 of the brain. It is still developing, but the idea is straightforward: People and AI agents should not have to reconstruct Zapier's latest thinking every time they start new work.

The team designed the system around three shared layers.

The first is shared context. The brain gives Zapier's people and AI agents access to current goals, product thinking, and brand guidance.

The second is shared stories. Paid media, social, events, and sales can adapt a message for their channels without inventing separate versions of the company's underlying truth.

The third is shared sprints. Once teams have the same context and story, they can take work to market together instead of passing it down a line.

Zapier is giving different teams and tools access to the same organizational memory without forcing every marketer into one tool or making every channel sound identical.

The Brain Meets Marketers Where They Already Work

For that common memory to matter, Zapier had to connect it to the places where the company's actual work happens.

The marketing brain lives in Zapier and can pull from sources such as Slack, Granola meeting notes, Zoom transcripts, Google Docs, and Coda. Marketers working in AI tools such as Cursor, Codex, or Claude can connect back to the brain instead of relying on an empty prompt and whatever company context they happen to remember.

That gives each campaign a different starting point.

"Every time you want to start a new campaign, you don't have to go rewrite the brief," Slagen says. "You don't have to go think about, 'Where are we starting from? What's the latest?' You just have the context, and you just go."

Then, as campaigns go live, the team can see where the brain missed something or an important input got dropped. The marketing itself becomes feedback for improving the brain.

The Simple Idea Became a Person-by-Person Build

When Zapier first discussed a shared marketing brain, the build did not sound that difficult.

Then the team got into the details: the tools each employee used, the way each person worked, the goals each team needed the system to understand, and all the places important context could live. The brain had to be built one person and one team at a time.

Zapier also learned that the brain needed a human owner.

Joe, the marketer who helped develop the "generational loss" idea, moved from product marketing into an agentic marketing role. He now owns the brain's maintenance. Slagen believes assigning at least one full-time marketing employee to the job is worth the investment.

"None of this would work without the maintenance piece," Slagen says.

Joe brings a technical background, product marketing experience, and years of Zapier context. He can understand what the brain should know, notice what it missed, and keep it aligned with reality.

Maintenance is part of the operating model, not a task that comes after the build. Without someone improving the context, a shared source of truth can become one more outdated source.

The Brain Is Already Showing Value

Zapier's internal V1 is too early for sweeping claims or a clean return-on-investment calculation. But Slagen can already see shared context changing how teams work together.

One example where this is playing out is in the relationship between brand and product marketing. Previously, the brand team could not keep up with every conversation between product marketers and the people building Zapier's products. As a result, brand concepts could drift from the latest product thinking.

But now, the marketing brain gives brand marketers a faster way to see what is current and identify which product marketing teams they need to consult.

The team is also seeing less generic work outputs from the AI tools connected to the brain. The tools can use current product direction and real examples of how a person speaks or writes instead of inventing an approximation.

"It has not erased AI slop completely, but it has significantly decreased the amount that I would've seen six months ago," Slagen says.

Still, better context does not guarantee good work. People need to review what AI produces. For instance, Slagen is comfortable having the system create email drafts and says he edits those drafts less than he used to. He is not comfortable letting it automatically send replies in his name.

The Brain Is Only as Good as the Context Behind It

Zapier's early build points to a larger lesson for companies trying to use AI across a team.

Giving people more AI tools does not solve the problem of fragmented information. It can make that problem more consequential. If every marketer and AI agent begins with a different version of the company’s most important information, they simply produce inconsistent work much faster.

Zapier's answer is to make organizational context reusable. Give people a shared starting point. Connect it to the tools and sources they already use. Assign someone to maintain it. Then use real work to expose what the system still does not know.

The brain does not eliminate judgment, handoffs, or adaptation. It is instead meant to keep all three grounded in the same current truth.

But if Zapier can keep the first copy sharp, the ninth one has a much better chance of staying on-brand.

Did you enjoy this transformation story? Go deeper on how AI is reshaping work and business with The Artificial Intelligence Show. Each week, we break down what matters in AI and what leaders should do about it.