There are two speeds inside every software company, and most teams only watch one of them.
The first is how fast you ship.
The second is how fast your users discover what you shipped.
For most of the last decade, those two ran close enough that nobody had to think about the gap between them. You shipped a feature every few weeks. People found it, eventually. Onboarding caught some of it. A webinar caught a little more. The lag was real, but it was small.
Then the build curve broke away from the discovery curve. And it hasn't looked back.
What changed on one side and not the other
AI coding tools changed how fast teams ship.
Cursor, Claude Code, Copilot. A PM with an idea on Monday can have it in front of users by Thursday. Teams that shipped monthly now ship weekly. Teams that shipped weekly now ship daily.
That is genuinely new, and it's good.
Here's the part nobody planned for: none of that touched the other curve.
Your users discover features at exactly the same speed they always have. A human still has to notice a new thing exists, understand why it matters to them, and change a habit they've had for a year. AI made your shipping 10x faster. It did nothing for the rate at which a busy person changes how they work.
So the two curves separated. One went vertical. The other stayed flat.
The space between them is the adoption gap. And every sprint, it gets a little wider.
The gap was already big before AI
This isn't a new problem that AI invented. AI just poured fuel on it.
Pendo has been measuring feature usage across thousands of products for years. Their benchmark: about 80% of features in a typical product are rarely or never used. That number barely moved from 2019 to 2024.
Userpilot found that the mean adoption of "core" features across B2B SaaS is 24.5%. Read that again. Three out of four users don't use the features you consider core.
When I was a Solution Architect at Mixpanel, I sat with hundreds of product teams and built the dashboards that showed them this. The reaction was almost always the same. Surprise, then a quiet kind of dread. They had shipped the thing. The thing worked. Nobody was using it.
Now take that already-broken discovery rate and put it next to a team shipping five times faster than before.
The features pile up. The discovery doesn't.
The old fix makes it worse
The instinct, when you notice the gap, is to buy something to close it.
A digital adoption platform. Build tours, write tooltips, set up checklists, define the segments and the rules. Point users at the new thing.
It works for a release or two. Then you ship again. The UI moves. The tour points at a button that isn't there anymore. Someone on your team has to go rebuild it.
So now you've added a second job that scales with your shipping speed. The faster you ship, the more tours break, the more maintenance you owe. You bought a tool to close the gap, and the tool's upkeep grows at exactly the rate the gap does.
That model was already strained when teams shipped monthly. At daily shipping, it snaps. No PM is going to hand-maintain tour content for a product that changes every day. The math doesn't work.
Why this hits the best companies hardest
Here's the uncomfortable part.
The gap is widest at the companies doing everything right.
The teams that adopted AI coding tools early. The product-led companies that ship fast on purpose because speed is their edge. The ones who most believe that shipping is winning.
Their build curve is the steepest in the market. Which means their discovery gap is the steepest too. The harder you push on shipping velocity, the more features you strand on the far side of a discovery rate that never sped up.
You can out-ship your own users. A lot of good teams already are.
The question worth asking
The two curves are not going back together by themselves. Shipping is only getting faster. Human discovery is not.
So the real question isn't "how do we build more tours faster." It's "why is a human building tours at all for a product that now changes faster than a human can document it."
If shipping became autonomous, discovery has to become autonomous too. Something that learns the product as fast as you change it, understands what a user is actually trying to do, and guides them to the feature they never found, without anyone writing a tour for it.
That's the bet we're making with Deway. Not a faster way to build tours. A system that makes building them unnecessary, so your discovery curve can finally keep pace with your shipping one.
But you don't need us to ask the question that matters this quarter.
If your team is shipping faster than ever, what is your plan for the users who haven't caught up to what you shipped six months ago?
Alon Binman is the co-founder of Deway (deway.ai), an AI-native autonomous adoption layer for SaaS products. Before Deway, Alon spent 15+ years at the intersection of product and customer success, including roles as a Product Manager, founder, data and product strategy consultant, and Senior Solution Architect at Mixpanel. You can reach Alon on LinkedIn.