activation is the make-or-break moment for any saas product. a user signs up. they land in your app. and then... something happens. either they reach their first "aha moment" in the next 10 minutes, or they close the tab and you've lost them. the data is brutal: the average saas product activates 30–40% of signups. the other 60–70% just leave.
the industry's answer has been product tours. onboarding checklists. tooltip carousels. explainer videos. none of it is working, and if you've ever been on the receiving end of a five-step "let us show you around!" tooltip sequence, you know why: they feel scripted, they break when you deploy a new design, and they can't answer "wait, what does this button do?" because they're not actually listening to you.
why tours fail
the fundamental flaw in tooltip-based product tours is that they're linear and scripted. the tour assumes the user will do things in a specific order. the user does not do things in a specific order. they click around, they explore, they take a wrong turn — and the tour either breaks or becomes irrelevant to where they actually are.
tooltip carousels have a 70–80% skip rate. most users click "skip tour" before the third step. the tour that took two weeks to build is ignored by the majority of the users it was built for.
the other problem is brittleness. every time you ship a design change, every time you move a button, rename a page, or restructure a nav, you break the tour. the CSS selector that pointed to "the blue button in the top right" now points to nothing. someone files an internal bug. someone else fixes it. two weeks of your engineering team's time, for a tour that 80% of users skip anyway.
finally: they can't handle questions. a user halfway through a tooltip sequence who wants to know "but what does workspace mean?" has nowhere to go. they can't ask. they can't deviate. they can only click "next" or "skip." this is not guidance. it's a slideshow.
flows that watch and adapt
aside's flows engine works differently. instead of placing tooltips on DOM elements, you define steps in plain language. each step has two parts: an instruction (what you want the user to do) and a detection condition (what the AI should look for to confirm they did it).
- step 1
instruction: "create your first workspace by clicking new workspace in the top left." · detection: "user has landed on workspace creation screen."
- step 2
instruction: "give your workspace a name and click create." · detection: "workspace has been created and user is in the workspace."
- step 3
instruction: "invite your first team member using the invite button in the top right." · detection: "invite modal is open or invitation has been sent."
the AI speaks each instruction aloud. then it watches the screen. if the user completes the step, it moves to the next one. if the user gets lost, it notices — "it looks like you clicked into account settings instead. let's go back to the main nav" — and adapts. if the user asks a question out loud (the widget is always listening), it answers before continuing.
no DOM selectors. no css classes. no brittle event listeners. the flow is defined in natural language, and it stays valid through design changes because the AI is looking at the screen the same way a human would.
visual highlights
for the most friction-heavy steps — forms, multi-step wizards, things that confuse every new user — aside highlights exactly where to click with visual rings on the user's actual page. human-in-the-loop: the user takes the action themselves and learns by doing.
the copilot says: "click the green 'create project' button at the top right" — and highlights the exact element. the user sees where to go and takes the action themselves. they learn by doing, not watching.
measuring activation
aside tracks three metrics that matter for onboarding, automatically:
flow completion rate
what % of users who start a flow reach the final step. identifies which step is causing drop-off.
time to first value
how long from session start to the activation moment. the number you want to drive down.
activation rate
before and after aside is enabled. this is the number that matters for the business case.
you'll also get step-by-step completion rates, so if 80% of users complete step 1 but only 40% make it through step 2, you know exactly where to focus. that data is available on day one — not after a quarter of A/B testing.
building your first flow
the flow editor in the aside dashboard is no-code. you don't write a single line of javascript. here's the full process:
- 01
name the flow. "new user onboarding" or "billing setup" or whatever the use case is.
- 02
set a trigger. options: first visit, specific URL match, or fire it via api from your own code (e.g., when a user completes signup).
- 03
add steps. each step is a plain-language instruction and an optional detection condition. drag to reorder.
- 04
click enable. the flow is live. no deployment, no code changes, no engineering ticket.
most teams build their first flow in under 20 minutes. the hardest part is deciding what your activation moment actually is — and that's worth thinking through regardless of what tooling you're using.
the question we always ask: what does a new user need to do to believe your product is worth paying for? that's your activation moment. build the flow around getting them there as fast as possible.