The companies winning the AI era ship software people genuinely love to use. Pendo surfaces where users and agents struggle and then helps you act on it, turning experience gaps into business results.
Every product manager has lived the same small heartbreak. You spend a quarter building a feature, you ship it, you wait - and almost nobody touches it. For years the only response was a shrug and a guess. In 2013, four engineers and product people in Raleigh, North Carolina decided the shrug was a business opportunity. They called the company Pendo, Latin for "value," and set out to answer the least glamorous question in software: what are people actually doing in here, and why won't they use the thing we built?
Thirteen years later, that question has grown into a platform used by roughly 14,000 businesses in 161 countries, a $2.6 billion valuation, and more than $200 million in annual recurring revenue. Pendo now sits inside the apps you use at work - watching, quietly, so the people who make those apps can watch too.
Drop a small snippet of Pendo code into a web or mobile app and it starts recording behavior without engineers having to instrument every button by hand. Product teams can then see which features get used, where people abandon a workflow, and how a new release changed the pattern. That is the analytics half. The second half is that Pendo can also talk back: build a tooltip, a walkthrough, or an onboarding checklist that appears inside the app - no engineering ticket required - to nudge a stuck user toward the part that works.
The combination is the whole idea. Most tools do one job: measure, or guide, or survey. Pendo's bet was that the value lives in the loop between them. See the drop-off, ask the user why, guide them past it, then measure whether the guide worked. Run that circle enough times and a product improves on evidence instead of opinion.
Pendo's customer list reads like a cross-section of the economy that runs on dashboards: Verizon, Morgan Stanley, LabCorp, Okta, Zendesk, OpenTable, Cardinal Health, RE/MAX. The company says it counts roughly 75 of the Fortune 500 and nearly half of the top 100 public software companies among its users. Some deploy Pendo on the apps they sell to the public; others point it inward, at the internal Salesforce or Workday screens their own employees have to survive every morning.
The problem Pendo attacks is old and expensive. Companies pour money into building features, then have almost no idea which ones justify the cost. Support teams field the same confused questions forever. Users churn without ever explaining why, because nobody asked them at the moment they were frustrated. Meanwhile the roadmap gets set by whoever argues loudest in the meeting.
Pendo's answer is to make the evidence cheap. Instead of a data engineer building a custom pipeline for every question, a product manager can open a dashboard and see the answer, then act on it the same afternoon with an in-app message. The pitch is not "more data." It is fewer meetings that end in a shrug.
Over the years the platform has grown a family of parts. Analytics tracks the behavior. Guides handle in-app onboarding and tooltips. Session Replay plays back a user's actual clicks and rage-clicks like security footage. Sentiment and NPS surveys ask people how they feel at the moment they act. More recently, Pendo Listen uses AI to synthesize scattered feedback into roadmap priorities, and Pendo Predict forecasts which accounts are drifting toward churn.
The newest and most talked-about addition points somewhere unexpected. Agent Analytics, which reached general availability in late 2025, measures how people interact with AI agents inside software - the same drop-off-and-confusion question, aimed at a new kind of user that types back. It is the feature that landed Pendo on Fast Company's 2026 Most Innovative Companies list.
The market Pendo plays in is crowded, but it is crowded with specialists. Amplitude, Mixpanel and Heap own hardcore product analytics. FullStory owns session replay. WalkMe and Whatfix own digital adoption and in-app guidance. Gainsight PX sits near customer success. Each is very good at its slice.
Amplitude, Mixpanel, Heap. Pendo does this.
WalkMe, Whatfix, Appcues. Pendo does this.
FullStory. Pendo does this.
Gainsight PX. Pendo does this.
Pendo's differentiator is the refusal to specialize. By bundling measurement, guidance and feedback into one product, it sells the loop rather than a single reading. A team does not have to stitch three vendors together to connect "users drop off here" to "so we showed them this" to "and it worked." For buyers tired of integration projects, one platform that closes the circle is an easier story than four best-in-class tools that do not talk.
Pendo is a straightforward B2B software subscription business. Pricing scales with monthly active users and product tier, with paid add-ons for things like data warehouse sync and deeper Salesforce and Segment integrations. There is a free tier for small apps, which acts as the front door - teams start with analytics on one product, then expand into guides, feedback and the enterprise features as they grow. In the fiscal year ending early 2024 the company crossed $200 million in ARR and reported its first quarter of positive cash generation, the sort of milestone that tends to precede an IPO conversation.
Pendo was started by Todd Olson, Eric Boduch, Rahul Jain and Erik Troan - alumni of Rally Software, Google, Cisco and Red Hat, where Troan had been an early engineer. Olson, the CEO, had run product at Rally and had two failed startups behind him before Pendo. He later wrote a book, "The Product-Led Organization," arguing that companies should put the product itself at the center of how they grow, rather than relying only on sales.
That thesis is not just marketing. It is why the company keeps buying capability instead of building it slowly. Four acquisitions in eighteen months - Zelta AI, Forwrd.ai, Mind the Product's Product Collective, and Chisel Labs - three of them AI-native, with a new R&D hub opened in Pune, India. The strategy is to fold new intelligence into the same loop as fast as the market moves.
Enterprise software has a house style: blue, serious, forgettable. Pendo went the other way. Its brand color is a hot pink the team nicknamed "Pank," and its annual gathering in Raleigh is billed not as a conference but as a "product festival," Pendomonium, complete with live music and a closing concert. It is an odd look for a company that sells to Morgan Stanley, and that is rather the point - a reminder that the software behind the dashboard was made by people who would rather not be boring.
Pendo's long-run bet is a simple sentence Olson has repeated for years: every company is becoming a software company. If that is true, then every company will eventually need to know whether its software is any good - and will not want to build the measuring apparatus itself. Pendo sells that apparatus. As AI makes software cheaper to generate and harder to trust, the company has repositioned around a blunt promise on its own homepage: the antidote to AI slop. Whether or not the tagline sticks, the underlying idea is durable. When anyone can produce software, the scarce skill becomes knowing if anyone can actually use it.