Cut my 6-step onboarding to 3. The realization that hurt: "personalization" was just delay.
Shipped phase 1 of the activation rebuild this week.
Old onboarding: 6 steps (name, role, work style, focus areas, peak
hours, weekend prefs). Ended on an empty calendar with "okay now
what?"
New onboarding: 3 steps. Name + role, then "what does today look
like?" (textarea, voice-enabled), then peak hours. Ends on a
planned day, fully laid out as time blocks, ready to execute. ~60
seconds from signup.
The second one is what the product actually is. The first one was
demographics collection sitting in front of the product.
Two calls that mattered more than they look:
-
Textarea defaults to today if signup is before 2pm local,
tomorrow if after. A 7am signup who plans tomorrow waits 17 hours
before experiencing the loop. That's a lost user. After 2pm
"what's left of today" feels sad rather than fresh, so the default
flips. Today/Tomorrow toggle either way. -
The textarea has a "speak it" button. Voice was always in the
app but buried in the dashboard. Surfacing it in the very first AI
prompt turns voice from "hidden feature" into "the way you talk to
Axtra Planner." New users discover it on signup — the only moment
they're paying attention.
Phase 1 of the 4-phase rebuild. Phase 0 was metered free (shipped
last week, results coming in 2 weeks). Phases 2-3 ship this week.
Curious how others have handled the "first 60 seconds" problem,
especially the demographics-gate temptation. Easy to convince
yourself collecting preferences = personalization, when really
it's just delay.
The painful part is most “personalization” in onboarding is just the founder postponing the first useful output.
Users do not care if the app knows their preferences yet.
They care whether it proves value before attention drops.
That is the real first-60-seconds test:
not “did they complete onboarding?”
did they get something usable fast enough to trust the loop.
The voice point is the stronger move here.
Not because voice is a feature.
Because it changes the mental model immediately.
The moment users speak instead of fill fields, the product stops feeling like setup and starts feeling like delegation.
That is the real shift.
Also: Axtra Planner is doing more work than the product UX right now.
It is clear enough, but still sounds like a feature wrapper.
If this keeps leaning into delegated planning instead of “AI planner,” something like Xevoa.com would age better.
The "trust the loop" framing is sharper than the "wow moment"
language I've been using. Wow is about delight; trust is about
whether the product proves it can take work off the user's plate.
That's the right metric for the next 14-day window — reframes what
I'm actually measuring.
On voice — yes, exactly. Surfacing it as the very first AI prompt
wasn't about showing off the feature, it was that typing-first
reinforces "I'm filling out a form" while speaking-first reinforces
"I'm telling someone what I want." Same data, completely different
mental model. I hadn't articulated it as cleanly as "delegation vs.
setup" but that's what's happening underneath.
The naming critique lands. "Planner" is feature-coded — tells the
user what the product IS, not what it DOES for them. The brand has
a few weeks of equity and a locked domain, so a rename isn't on the
table this quarter. But the substance of the point is going to
shape how I write copy. Less "planner that builds your day," more
"tell it your day, it does the planning." That's the orientation
shift you're pointing at, even if the URL doesn't change.
Exactly — “delegation vs setup” is the real product shift.
And that’s why the naming issue may show up faster than expected.
If the product experience says:
“tell it what you want, it handles the planning”
but the name still says:
“planner”
then every first impression starts one step behind the actual product.
Copy can help, but the name still sets the first mental model.
So I’d be careful treating it as a next-quarter problem if the next 14 days are about proving trust in the loop.
If users are already moving from setup to delegation, the brand should not keep pulling them back into “AI planner” territory.