The product tour was never onboarding. It was a broadcast, the same path, narrated at the same pace, to every user regardless of who they are or why they came. Real onboarding has always been a conversation: questions answered, context absorbed, pace adjusted. We built tours because conversations didn't scale. That excuse just expired.
What a tour assumes, and why every assumption is wrong
A tour assumes the user's goal matches the authored path (it doesn't; an admin and an analyst arrived for different reasons). It assumes their questions will occur in the authored order (questions are famously rude about sequencing). It assumes attention is available for the tour's full length right now (the user has a meeting in seven minutes). And it assumes silence is comprehension, the most expensive assumption in SaaS, because a tour cannot hear confusion. It just advances to step four.
What conversations do that broadcasts can't
Watch a great onboarding specialist for ten minutes and count the adjustments: they answer the interruption, notice the hesitation before the click, skip the module this customer clearly doesn't need, circle back to the thing that didn't land. None of this is content. All of it is response, and response requires perception (seeing where the user actually is) plus adaptation (changing the plan without losing it). That last clause matters: a good conversation isn't planless; it's a plan held by something that can bend. This is exactly the agenda-driven session model, the difference between software that presents and software that listens.
The false economy we all accepted
Tours won because the alternative was headcount. Fine, that was the actual constraint, and pretending the tour was pedagogically equivalent was the industry's polite fiction. But notice what we normalized: the moment of highest customer intent and highest churn risk, handled by our only fully deaf medium. Conversation-grade onboarding at broadcast economics is now a real option, voice, screen-aware, adaptive, multi-session (disclosure: it's what we build at Skippr, and others will build it too, because the fiction was never load-bearing). The question for teams is no longer "can we afford to converse?" It's "what's our excuse for broadcasting?"
A test you can run this week
Take your tour's biggest drop-off step. Sit five new users in front of it and let them ask one question out loud at that moment. Write the five questions down. Your tour answers none of them, that's why they left. A conversation answers all five, in the order they were asked. That's the whole argument, in one usability session.
Watch it onboard someone
Onboarding either happens on the user's own screen or it does not happen. Fifteen minutes is enough to see which one you are buying.