There's a moment every company hits and almost no one talks about it honestly.

Revenue is moving. Product works well enough. The early panic has settled into something more like a grind. The team is tired in the good way, the we've-been-through-something way. And then quietly… the meetings change.

Someone brings a deck about an adjacent opportunity. A new hire, fresh from somewhere bigger, starts asking why you're not doing what they did at their last place. A potential partner comes in with a white-label proposal that makes the numbers look really good for about forty minutes. The founder who has been staring at the same dashboard for eighteen months and having the same conversation with the same thirty customers starts feeling the itch.

This is the moment. Not the product failure, not the bad hire, not the missed quarter. This is where most companies actually lose, long before anything shows up on a spreadsheet. The itch gets scratched. The meeting gets booked. And the thing that was quietly becoming something real starts quietly becoming several things that never quite do.

88% resolved. 22% stayed loyal. What went wrong?

That's the AI paradox hiding in your CX stack. Tickets close. Customers leave. And most teams don't see it coming because they're measuring the wrong things.

Efficiency metrics look great on paper. Handle time down. Containment rate up. But customer loyalty? That's a different story — and it's one your current dashboards probably aren't telling you.

Gladly's 2026 Customer Expectations Report surveyed thousands of real consumers to find out exactly where AI-powered service breaks trust, and what separates the platforms that drive retention from the ones that quietly erode it.

If you're architecting the CX stack, this is the data you need to build it right. Not just fast. Not just cheap. Built to last.

See the data

A flood carries more water than a river. Way more. It moves with tremendous visible energy, spreads across everything, touches every surface, covers miles of ground in hours. Then it's gone. Absorbed into soil and sun. The landscape looks almost exactly like it did before.

A river moves a fraction of that volume. But always in the same direction, against the same resistance, wearing down the same rock. A river moving for long enough cuts a canyon.

Most companies die floods. They move everywhere at once and leave nothing permanent behind. And the brutal part is that it feels like progress the whole time. New products, new markets, new metrics lighting up on the dashboard. The sensation of momentum is almost indistinguishable from actual momentum. For a while the numbers cooperate. Then they don't. By the time the damage is visible the cause has been completely buried under the language of bad luck and shifting markets.

In 2015, Notion was four people in Kyoto running out of money.

Ivan Zhao had rebuilt the product from scratch once already. When the rebuild didn't work the way he wanted it to, he rebuilt it again. The team moved to Japan partly because it was cheaper, partly to get away from the noise. There were real conversations about whether to keep going. The product wasn't working. The runway was short. Most people would have pivoted, expanded the thesis, found an adjacent use case to chase.

They kept doing the same thing. Made it better. Kept doing it again.

Shipped the new version in 2018. Posted it on Product Hunt on a Tuesday, which is not when you're supposed to post on Product Hunt if you're being strategic about it. It went viral anyway. Not press release viral… actual viral. Servers crashing, inbox breaking, people sharing it because they wanted their friends to have it. The kind of response you can't manufacture with a launch strategy.

Series A in 2019. Sequoia led. Valuation: $800 million. Team at that point was still around ten people. Ten people, $800 million valuation, one product that was still basically just a really thoughtfully designed notes and docs tool.

By 2021, another $275 million, $10 billion valuation, around 4 million users. And still… a notes tool. Databases had launched. But the core thing, the reason people came, was still the same thing it had always been. Blocks. Documents. The feeling that the software respected how you actually think.

The mobile app took forever. Users complained constantly. It was genuinely bad for a long time and everyone knew it. The criticism was fair. Notion shipped it when it was right, not when the noise got loud enough. Which is almost never the decision that gets made.

Meanwhile Evernote.

Evernote raised $290 million. Two hundred and ninety million dollars. Had 225 million registered users at peak, which is a number most software companies would retire on. People called it the category definer. Phil Libin, the CEO, was giving keynotes about how Evernote was going to be a hundred-year company.

Then the meetings changed.

They shipped a food diary app. A physical notebook with Moleskine. A business card scanner. A presentation tool. Evernote Market, which sold physical goods… bags, laptop sleeves, Post-it notes. Actual branded socks. Evernote socks. At some point someone in that building sat in a room and said we should sell socks and enough people agreed that it happened.

The note-taking product, which was never quite finished, which always had sync issues and a cluttered interface that never got resolved because the energy was going somewhere else, started losing the people who had built their workflows around it. Not dramatically. Slowly. The way trust erodes when a product stops feeling like it's being made by people who care about it.

By 2022/23 they sold to Bending Spoons, an Italian app company that buys distressed software. Terms undisclosed. For a company that raised $290 million and had two hundred and twenty five million users. That's not a bad outcome. That's a complete destruction of value, built incrementally, one sensible-seeming decision at a time.

Michael Porter wrote about this in the 1990s and it still holds up better than most frameworks people are running around citing. The argument was simple. Companies don't fail because they pick the wrong strategy. They fail because they try to run two strategies at once. Cost leadership and differentiation. Broad market and niche. You end up stuck in the middle, which Porter called exactly that… stuck in the middle. Mediocre at everything, excellent at nothing, getting beaten on both ends by people who made a choice and stayed with it.

Evernote was stuck in the middle. A power tool for serious note-takers that kept making decisions aimed at casual users. A consumer product that kept trying to sell to enterprise. A software company that sold socks. None of the strategies were wrong exactly. Running all of them simultaneously made each one worse.

Clayton Christensen had a related framework, less famous than the innovator's dilemma but more useful here. He called it RPV. Resources, processes, values. The idea was that what a company can actually do is determined not just by money or headcount but by the processes it's developed over time and what it prioritizes when it has to make a tradeoff.

The problem with expansion is almost never resources. You raise a round and now technically you have the money. The problem is processes and values don't transfer. The judgment that made product decisions feel obvious in one context makes them hard in another. You're not building a new product. You're trying to build a new company inside the old one, same people, same time, same finite amount of attention.

Notion's value, the thing it chose when it had to make a tradeoff, was craft over speed. Every time, for years, while the criticism was loud and the mobile app was bad and the competitors were shipping faster. That consistency is what turned a notes app into something people trusted completely. By the time Notion launched databases and project management the users didn't need a feature comparison. They already believed the taste was better. That belief was the product of years of watching Notion refuse to ship things that weren't right yet. You can't buy it. You can only accumulate it by staying with something long enough that people run out of reasons to doubt you.

Here's the honest version of the advice that usually gets softened into something more comfortable.

Most of what feels like strategic thinking in the early stages is restlessness dressed up in the language of strategy. The second product exists because the first product is hard and the founder is tired of the same problem. The partnership conversation exists because organic growth is slow and a deal feels like a shortcut. The adjacent market exists because the current path requires patience the room doesn't have.

The founders who build something real are not smarter about markets. They're just more honest about their own psychology. They see the itch for what it is and keep working anyway.

Evernote saw the itch and called it a roadmap.