Zapier becomes integration infrastructure
Senior executive at no-code startup on the rise of native integrations
The strategic problem is that Zapier wins when users leave their app to automate, but the market is shifting toward automation that stays inside the app. In practice, the best integrations are usually the top 10 or so workflows that lots of customers use every day, like Slack alerts, CRM syncs, and email triggers. Those are exactly the ones SaaS companies now have the strongest incentive to make feel native, which pushes Zapier toward supplying the infrastructure behind the scenes instead of owning the user interface.
-
Embedded changes who the customer is. Instead of selling a business user on building a Zap in Zapier, the product is sold to a software company that wants its own users to click a native looking Connect button, authenticate once, and run automations without ever leaving the product. That is the lane Tray.io, Alloy, and Paragon are built for.
-
The reason this matters economically is that Zapier’s strongest workflows are also the easiest ones for partners to internalize. The interview points to a world where SaaS companies keep the head of demand, the few integrations most users need, and leave Zapier with the long tail of edge cases. That is still useful, but it is a weaker position with less pricing power.
-
Zapier still has real advantages, scale, distribution, and workflow data. It reached about $310M in ARR by 2023, versus roughly $70M for Tray.io, and its installed base and app coverage remain far larger. But those strengths matter most if they are turned into rails that other products build on, not just a destination site users visit separately.
The next phase of the market is likely a split model. SaaS products will own their most important integrations in product, and specialized platforms will power that experience underneath. If Zapier keeps moving closer to embedded and orchestration, it can stay central to automation. If not, it risks becoming the fallback option for everything a product team chose not to build itself.