I've been writing about AI and automation for Zapier for the last two years. In that time, I've seen Zapier ship one useful AI tool after another: some for automating securely with AI, and others for bringing secure automation into your AI tool of choice.
Keeping up with these releases has been part of my job, but it doesn't have to be yours. Whether you're new to Zapier or a longtime automator, I'll introduce you to the whole AI lineup, so you're fully prepared to plan your next build.
Jump ahead
Add AI steps to your Zap workflows
A Zap—that's our name for an automated workflow—completes tasks for you without your manual input. Not every Zap needs AI. But when some part of your workflow calls for reasoning or creative judgment, an AI step can handle it.
That's what AI by Zapier is for. It's our built-in tool for adding AI steps to your workflows, powered by your model of choice from OpenAI, Anthropic, and Google. And it puts you back in control over your workflow cost. You no longer have to run the whole nine yards of a workflow through an AI model, which can get expensive. You can just drop AI steps into a Zap exactly where they're needed.
For example, let's say that when a new row is created in a Google Sheet, you want an automation to ping a teammate in Slack. You can build a Zap for that, then add an AI by Zapier step in between the Google Sheets trigger and the Slack action. That way, instead of just sending your teammates a barebones alert about a new row, you can have the AI step summarize the row's contents and weave that into the notification, providing helpful context. And because that middle step is the only one that runs through a model, it's the only part of the Zap you pay AI prices for.
AI steps don't have to work from scratch, either. You can ground them in your company knowledge, like connected docs from Google Drive, Notion, Confluence, or wherever your context lives. You can also make an AI step agentic by letting it call app actions at runtime—like a web search or a lookup in your CRM—to reason through a task in multiple moves.
And AI by Zapier isn't your only option for adding AI steps. Nearly 700 AI apps integrate directly with Zapier, so you can hook up whatever you already use and love to your workflows.
Store data in AI-powered tables
In between steps in a Zap workflow, you might need to store information, like a list of leads, a queue of drafts waiting for approval, or a log of processed refunds. A spreadsheet can technically store that. But it's not designed for heavy, rapid writes. If your Zap runs hundreds of times per day, you might hit rate limits that slow your workflow down. Plus, pulling data from a spreadsheet into another app can be clunky, potentially requiring hacks like lookup formulas or extra search steps.
Zapier Tables fixes that. It's a database built into the platform, and it's included on every plan—including the free one—so everyone can easily and securely route their data across more than 9,000 apps.
Two AI features live inside Tables: AI fields and AI enrichment.
AI fields run a prompt you write against every row of a particular column, including rows that already exist. For example, say you keep inbound leads in a table. You could set up an AI field with a prompt that drafts a short intro email for each lead, pulling their name, company, and product interest from the rest of the table. If a referenced field changes later, the AI field regenerates on its own, so the output never goes stale.Â
AI enrichment, on the other hand, is a table-level feature. You write a single prompt and tell it which existing columns to fill, and it'll populate multiple columns of a row. (It won't backfill existing rows, though.) For example, you could write a prompt like "Given the company name, generate the likely industry, headcount range, and headquarters city." Then, whenever you add a row with just a company name, those three columns will populate themselves based on what the AI knows.
Install Zapier in your AI agent harness
Do you and your team prefer to work inside AI tools like ChatGPT, Claude, or Cursor? In that case, every app you connect to your AI needs to be governed—otherwise, you're exposing your private company data to significant security risks. With Zapier, you can connect your AI to our full app ecosystem securely using one of three install paths:
Zapier MCP, which hands your AI chatbot a curated menu of 30,000+ pre-built actions
The Zapier SDK, which lets you build inside your coding agent with loops, conditionals, error handling, and calls that reach past the pre-built catalog
The Zapier CLI, which lets you explore apps, discover what they can do, and run actions from your terminal
To learn more about installation and the differences between each path, check out our feature guide.
Note: The Zapier SDK is currently in open beta. Although pre-built actions are governed, direct calls to raw API endpoints are not.
Other intelligent building tools across Zapier
Okay, so you know which products can pull AI into your automation (or vice versa). But we also have AI tools that help you bring your automation ideas to life.Â
For example, if you need help visualizing complex processes, there's Zapier Canvas. Canvas lets you map out entire workflows with AI-powered diagrams, so you can identify bottlenecks and keep everyone aligned on how your team operates.
You also have access to some huge AI power-ups across the platform:
Located in the member home page and inside every product, Zapier Copilot can build complete, cross-product systems with just a simple, plain-language prompt. It's contextually aware, too, meaning it always gives you the most up-to-date, helpful suggestions for building workflows.
To customize your Zap workflows, create code steps by describing what you want in plain language to auto-generate JavaScript or Python—then use Formatter with AI to get intelligent suggestions to format data correctly between apps. You can even access the Zapier SDK inside a code step. That lets you call any connected apps from your code without hard-coding API keys or building a private integration.
Don't see an action you need? Custom Actions lets you design and share team-specific features using AI. And no coding knowledge is required to use it.
If your Zap fails, AI troubleshooting will automatically give you plain-language explanations and solutions, so you can resolve issues quickly.
Build an extra safety net into your automations with AI Guardrails by Zapier. Add it as a step to a Zap to screen content for personally identifiable information, prompt injection attempts, toxic language, and negative sentiment—before any of that reaches your business systems.
Put AI to work across your business, all in one platform
Building with AI can feel risky when you've got sensitive data to protect. But with Zapier, you can put AI to work without compromising your governance, security, and compliance standards. If you're ready to transform your operations, see what AI at Zapier can do for you today.
This article was originally published in May 2024 by Elena Alston and was most recently updated in August 2026.
