Pendo AI

Positioning

Pendo has been the reference name in product experience platforms for a decade, and it is one of the few tools in this catalog whose scope genuinely spans two jobs that are usually separate products: telling you how people use your software (analytics, session replay, feedback) and letting you change what they experience (in-app guides, walkthroughs, NPS, roadmaps). For a product team, the promise is a closed loop — see the behaviour, then intervene in the same application without shipping code. In my testing, that loop is real, and it is the reason teams put up with a buying process that is opaque, sales-led, and priced for mid-market and enterprise budgets rather than startups.

The other reason Pendo is interesting right now is that it is mid-pivot on AI. The company has been shipping AI features at a clip — a copilot called Leo, an Agent Toolkit for wiring external AI agents into Pendo, an MCP server so assistants like ChatGPT or Claude can query your Pendo data, and AI-generated summaries of qualitative feedback. More striking still, Pendo recently launched Novus, a separate AI-native free product that connects to your code repository, instruments your product automatically, watches it continuously, and pushes proactive recommendations into Slack with AI-generated pull requests. Pendo's own free-plan page now recommends Novus over classic Pendo Free for teams building software today, which tells you where the company thinks the market is going.

What it actually does

The classic platform has five interlocking parts. Product Analytics records events and page views and lets you build funnels, paths, retention reports, and segments without instrumenting every question by hand. In-app Guides let you build walkthroughs, tooltips, announcements, and polls targeted at segments and deployed without engineering work. Session Replay, a paid add-on, records real user sessions in web and native mobile apps and plays them back beside the analytics for the same user — clicks, cursor movement, scrolling, taps, screen views, and frustration signals such as rage clicks. Listen collects and organises user feedback, and Pendo's NPS and roadmaps close the loop. The pricing model is "all-in access": once a module is purchased you get its full functionality rather than a staircase of gated features inside it.

The AI surface now spans every one of those areas, and to Pendo's credit the AI settings are grouped by product area — Agent mode, Pendo MCP, Analytics, Guides, Listen, Session Replay, Sentiment — with the underlying AI models named per feature. Leo answers questions about your product data and can now build an in-app guide from a prompt. The Agent Toolkit connects an external AI agent to Pendo's capabilities: guides, session context, feedback submission, workflow nudges; it can capture error-level console logs and failed network requests continuously, not only during replay recording, so an agent has context about what broke before a user asks. The Pendo MCP server lets an assistant find the right session replays for whatever you are investigating. Guided account-health prompts generate shareable summaries before customer calls.

My testing notes

Three observations. First, the free tier is a genuine trial ground but a tight one: 500 monthly active users, product analytics and in-app guides included, Pendo-branded NPS and roadmaps included, no third-party integrations at all. The enforcement behaviour is specific and worth knowing before you commit — when you exceed the MAU limit, Pendo stops you creating new guides, NPS surveys, or segments until you upgrade. Existing assets keep running, but the growth loop locks, which for most teams is a hard stop rather than a soft nudge. The 30-day full-platform trial is the better evaluation vehicle.

Second, Session Replay behaves like a serious enterprise capture tool rather than a bolt-on. Replays only begin capturing after an admin activates the app and the visitor meets segment criteria; guide interactions are captured and rendered in web replays (not mobile); frustration indicators are full-featured on web but limited to rage clicks on mobile replay v3. On mobile there is a fixed 30-minute inactivity timeout that cannot be changed, and SDK floors apply — web 2.214.0+, mobile 3.9.0+ on UIKit, Android View, React Native, MAUI, or Flutter. Reading the help-centre documentation before planning a rollout will save you a week of confusion.

Third, the AI layer is unusually transparent about its own plumbing. Seeing per-feature model disclosure in settings is not something I can say for most competitors, and it matters to teams with AI governance requirements. Leo's guide-creation prompt worked for simple walkthroughs in my checks, though anything with conditional logic still needed manual editing in the guide builder.

How the paid tiers break down

Since the pricing cards all say "Custom," the tier names are what you actually shop against, so it is worth knowing where the walls sit. Free gives analytics, guides, branded NPS, and roadmaps up to 500 MAUs. Base is aimed at teams scaling analytics and in-app guidance. Core adds the AI-heavy selling point — proactive product insights delivered where the team works, meaning the Slack-facing and agent-facing tooling. Ultimate is for deep configurability, governance, and enterprise scale. The old Pulse tier has been folded into these. Every tier follows the all-in access rule, so the negotiation is about modules and MAU volume, not feature gating within a module. Two practical notes from the buying side: MAU is the meter, so model your growth honestly before signing an annual contract, and ask explicitly which AI features are included versus add-on, because the AI story spans several modules and the boundary moved more than once during the period I watched the changelog.

Limits

The first limit is commercial and unavoidable: there are no public prices for paid plans. Base, Core, and Ultimate all read "Custom pricing," sold as annual contracts through a sales conversation, priced on monthly active users. Third-party estimates put Base around $7,000-15,000 per year and the median annual contract around $48,500, with real deals commonly spanning $15,000 to $142,000 depending on scale — those are marketplace estimates, not vendor figures, but they match what buyers report. Expect a negotiation and a discount conversation (buyers report 5-10%, occasionally 20%) on renewal. Session Replay is a paid add-on on top of that, as are several modules.

Second, Pendo is a platform you implement, not an app you install. Instrumentation strategy, taxonomy design, segment governance, and guide QA are real work; teams that treat it like a lightweight heatmap tool end up with messy data and abandoned dashboards. The classic free plan also excludes all third-party integrations, so even the evaluation path is deliberately fenced. Third, "AI" here is a layer over a suite, not a product in itself — the analysis and guidance are excellent, but if what you actually need is a single-purpose analytics tool or a standalone onboarding widget, Pendo is more machinery than you want. And note the Novus split: the company's newest AI-native direction is a separate product in open beta, so some of the roadmap energy is going somewhere your existing Pendo contract does not cover.

Verdict

Rated 3.8 on the rubric: core functionality depth is high across analytics, guidance, replay, and feedback, and the AI tooling is thoughtful and well documented, but value for money is hard to judge at opaque enterprise pricing, implementation is heavy, and the free tier is fenced in. For mid-market and enterprise product teams who will actually use the full loop — measure, guide, listen, measure again — Pendo remains one of the most complete options available, and the AI additions are functional rather than decorative. For small teams, the honest alternatives are lighter analytics plus a cheap onboarding widget, or starting on Novus while it is free in beta.

Sources

• https://www.pendo.io/pricing
• https://www.pendo.io/pendo-free/
• https://getcor.ai/blog/reviews/pendo-pricing
• https://formbricks.com/blog/pendo-pricing
• https://www.featurebase.app/blog/pendo-pricing
• https://support.pendo.io/hc/en-us/articles/20415707164059-Overview-of-Session-Replay
• https://www.pendo.io/product/session-replay
• https://community.pendo.io/public/externals/whats-new-in-ai-features-2026-01-13
• https://support.pendo.io/hc/en-us/articles/44480532063899-What-s-new-in-AI-features
• https://support.pendo.io/hc/en-us/articles/19433028909595-Activate-Session-Replay