"I put AI to work."

Zapier has built a platform that changes who inside a company gets to build. Most tools promise features; Zapier has quietly changed the cast. In June 2025, Wade Foster announced a milestone few companies could claim: Zapier now runs more AI agents than it has employees. The more telling detail is who made them — support agents, recruiters, ops leads. People whose job titles say nothing about software.

When Zapier works, it gives a company two things. The first is builders: people who thought AI wasn't for them, designing and shipping it into their own work — without waiting on another department. That's the gift in every adoption number Zapier publishes, in the hackathons, in the rollout story that took internal AI usage from 10% to near-universal.

The second is coherence: scattered builds becoming how the company actually works. One workflow is a win; a thousand workflows are either a machine or a mess. That's the gift that turns "some of our people automated some tasks" into "our company operates differently now."

The first gift is told everywhere, in feature language. The second is told almost nowhere a customer walks. The distance between those two stories is where the opportunity lives.

Right now, the builders gift arrives in different voices depending on the surface. Take the AI page:

The second gift, coherence, barely gets told at all — even though Zapier's own research names the fear it answers. Executives are not asking "will my people automate a task." They are asking whether a hundred scattered AI experiments will ever add up to a governed, working whole — or remain pilots forever. Zapier has published the data on exactly this anxiety. It has published the four pillars it uses to resolve it. It has a CEO who tells the internal transformation story — the code red, the hackathon, the fluency bar — with unusual candor.

But that telling happens on podcasts, in playbooks, in cohort programs and long-form posts for transformation leaders who already opted in. On the surfaces where a buyer actually meets the brand, the promise stays at the task level. Which means the full sentence Zapier could be giving companies only ever gets its first half said out loud:

The second half is what the executive sponsor needs in the room where the budget is defended. Zapier even has the custodial line for it, spoken by its own CTO and repeated inside its talent organization: "You can delegate the task, but you can't delegate the accountability." That sentence is the machine's conscience — the reason scattered builds cohere instead of sprawl. It is true, it is public, and it lives in hiring materials, not on a single surface a customer sees before the deal.

The opportunity is to tell coherence on purpose, on the surfaces where the funnel actually runs — so the buyer doesn't have to infer, from a features page, whether a thousand builders become a machine or a mess.

That's the whole reading. One platform, two gifts, and a brand that tells the first in feature language and the second in rooms the funnel doesn't enter.

My working thesis, the thing I've spent the last year on, is helping brands say clearly what they actually give people. This is not about features. It's about the change in someone's work and self-story, told in a voice that persists across every place they meet you. Zapier is the rare company where both gifts are already real — the builders exist, the machine exists, the language for both exists inside its own walls. The narrative just hasn't caught up to what the product and culture have already made possible.

If any of this is useful, I'd love to talk.