Josh Royse spends a lot of time helping teams get to the moment Pendo is actually for: when you start to understand how your product is being used. As one of Pendo’s Customer Engineers, he's the person in the room when that transition happens, and he'll tell you it's always started in the same place.

Before any of it clicks into place, you have to describe your product to Pendo. Page by page, feature by feature, using proper naming conventions, product areas, and descriptions. It's just what meaningful product analytics has always required across every tool in the category.

"For a decent-sized app, tagging used to mean two weeks," Josh said. "Now I'm looking at a couple of days."

That’s thanks to AI tagging, now available in open beta for all Pendo customers. And the fastest way to understand what it does is to hear what happened when Josh first used it with a customer who was evaluating Pendo.

He skipped the sanitized demo environment and enabled AI tagging directly on their subscription, tagging their actual product in the session. The reaction from their team was immediate: "THIS is what I want AI doing for our lives."

That’s what product people who've spent real time on instrumentation suddenly saw: it collapsed into an hour or two of work.

What AI tagging makes possible

The capability matters differently depending on where you are with Pendo.

If you're new, that means you can have meaningful analytics coverage within your first week, before you've written a single CSS selector.

If your product ships constantly, AI tagging keeps up. After a release, new pages and features surface automatically in the suggestions panel, ready for you to review and accept. The maintenance backlog that used to follow every sprint no longer accumulates the way it did.

If you've been on Pendo for years and your tags have quietly degraded every time something ships, AI tagging surfaces what's stale, what's broken, and what's missing, along with suggestions for everything that's new. The coverage that your analytics, guides, and AI all depend on stays current without the overhead.

How AI tagging works in Pendo

When you open the designer, AI tagging launches by default, reads your application's routing configuration and DOM, and streams suggestions into the panel as they're ready.

  • Tagging suggestions include a proposed name, rule, description, product area, and page-to-feature link.
  • Hover over any element in your application, and the panel jumps straight to the suggestion for that element, so reviewing what you're looking at takes a second rather than a search.
  • Confidence indicators flag where the AI is certain and where a human call adds value.
  • Sensitive data is sanitized client-side before anything leaves your application.
  • Nothing is applied until you explicitly accept it. Your existing tags are never modified unless you approve the change. You stay in control of what goes into your subscription.

Accurate tag coverage matters far beyond cleaner dashboards

Tags are the foundation on which everything else in Pendo sits: guide targeting, segments, Leo's proactive signals, and MCP-powered workflows. As AI capabilities have expanded what's possible, the demand on that foundation has grown with them.

MCP now lets teams build a much larger volume of reports and analyses than anyone could produce manually, pulling structured product data into automated workflows at a scale that wasn't practical before.

But the output is only as good as the input underneath it: more AI-powered reporting means you need more named, structured coverage to draw from. AI tagging is how you build and maintain that input without the overhead growing alongside it.

When coverage is incomplete, guides miss their audience, segments don't reflect reality, and every AI capability (including the ones doing work you'd never have time to do yourself) operates on a partial picture.

Why Pendo’s approach to AI tagging is different

Two categories of tools have tried to solve this before, and neither fixes the problem:

  1. Autocapture tools record everything by default, which may sound like the answer until you're looking at hundreds of unnamed, raw events with no structure, no hierarchy, and no product areas. You have coverage, but you still have to make sense of it before any of it is useful for analytics, guide targeting, or segmentation. The investment moves; it doesn't disappear.
  2. Other tools use AI to keep in-app guidance working when a UI changes, recognizing elements automatically so walkthroughs don't break when a button moves. That's a meaningful capability, but it's a guidance maintenance tool. It doesn't identify or name pages and features, doesn't label your event stream, and doesn't produce the structured product coverage that analytics and guide targeting depend on. Customers on those platforms still manage their product structure manually.

AI tagging fixes both issues at once: it identifies what's worth naming, names it with a hierarchy and product areas, and makes it immediately usable across analytics, guides, and AI, without a separate manual step.

Complete coverage of the right things

This is where Josh is direct about something important: AI tagging isn't designed to tag your whole application automatically, and that's a deliberate choice.

"I created 300 tag suggestions in one session," he says. "But I still had to decide which of those 300 actually matter. An 'About Us' page probably doesn’t need to be tagged. You still need to use your head."

You know your product and your use cases better than any AI does. AI tagging removes the physical legwork, and thinking about what matters stays with you. The goal is complete coverage of the right things, not coverage of everything.

Get started today

AI tagging is in open beta for all Pendo users. A subscription admin can enable it in Settings > AI Access > AI tagging. From there, go to Product > Features, select Tag Features, and AI tagging opens by default in the designer.

For a full walkthrough of the setup, naming conventions, and selector configuration, see the AI tagging overview and Tag with AI in the help center.