More of your customers are working inside AI tools now, not switching between apps. If they can't reach your product from inside Claude or ChatGPT, they're going to use something they can.
Zapier's co-founder and CTO, Bryan Helmig, and Zapier Director of Product Management, Lisa Chapello, joined our first partner webinar of the year—Building Smarter with Zapier: AI Updates, New Ways to Partner, and What's Coming in 2026—to talk through what that actually means for software companies. Here's what stood out.
Your competitive edge isn't going away—and neither is software
There's a lot of "SaaS apocalypse" energy out there right now. AI valuations swinging wildly, vibe coding making headlines, and genuine questions about whether software products are still worth building. Bryan's take: that's the wrong read.
"We're going to have more software as a result of AI, not less," he said. "What changes is the bar. Cheap, mediocre software is easy to build now. What people are willing to pay for is rising. That's actually an opportunity if you're building something worth paying for."
The companies that stay in the workflow aren't the ones who wait.
Your product may already be accessible inside AI tools
Matt Marum, who leads Zapier's partner activation work, hears a version of the same thing from vendors every week: "A lot of vendors that have Zapier integrations aren't necessarily aware of what we've done with Zapier MCP—and they might not actually know that their products are already accessible today, even if they haven't implemented MCP themselves."
If you have a Zapier integration, your app's actions—creating records, sending messages, updating contacts—can already be called by AI tools like Claude, ChatGPT, and Cursor through Zapier's MCP support.Â
MCP is a protocol, sometimes described as USB for AI, that lets AI tools discover and call external tools. Zapier supports it across 9,000+ apps and 30,000+ pre-built actions, which means the integration work your team already did is now available to AI. Your users can reach your product from their favorite AI tool without ever switching tabs.
So, the question isn't whether to build for AI. It's whether your existing integration is built well enough to show up when AI is doing the choosing.
Connect with our team to see where your integration stands. Matt Marum leads our partner activation work and can walk you through it.
How to make sure AI picks your app
In a traditional integration, a user picks your app. In an AI-driven workflow, the model picks it—or doesn't. That rewards different things than traditional integration design.
What helps agents find and use your product:
Clear action names: Actions should be specific enough that an AI understands what the action does without extra context.
Focused scopes: Create actions that do one thing well, not five things loosely.
Minimal required fields: Fewer required parameters means the AI can call your action without guessing.
Good defaults: Build sensible fallbacks so the model doesn't have to infer values to do something simple.
Solid documentation: Agents read your API docs; make sure they're accurate and current.
Small fixes—renaming an action, cleaning up required fields, adding a search capability—can make a surprisingly big difference in how often your integration shows up.
Make AI-powered automation native to your product
Being callable from AI tools is one thing. Making automation a native part of your product experience is another—and it's where the biggest revenue and retention gains are.
That's what Powered by Zapier is built for. Instead of your users leaving your product to trigger automations in a separate tool, you embed Zapier's automation capabilities directly inside your product. Your users build and run workflows without ever seeing a Zapier login screen. And because you can cover the cost of their automation tasks, they get it for free—which removes one of the biggest friction points in getting users to actually adopt automation.
TikTok and Typeform are already doing this. They embed Zapier inside their products and sponsor their users' automation tasks, so the experience feels native, not bolted on. For their users, it's just a feature. For TikTok and Typeform, it's a retention lever.
As more of your users start building AI-assisted workflows, the ones who can do that without leaving your product are going to stick. Embedded automation—especially with MCP support—means those workflows happen inside your experience, not someone else's.
Powered by Zapier is for companies that want to embed automation in their product. You can cover the cost of tasks for your users so they get automation for free without ever leaving your product. Learn more →
The most important question to ask about your integration
Getting into the workflow used to mean: do we have a Zapier integration? Most teams reading this can check that box. The new question is whether you're built for the workflows that are actually growing.
The reframe, as Lisa put it: "Is our integration agent-ready?"
Look at your user data. There's probably a small group—maybe 2%—already using your product in ways that reflect where things are headed: AI-assisted, cross-app, automation-heavy. Find them. Figure out what they're doing. Then figure out how to bring that to the rest of your users as fast as possible.
