Generate guided learning paths from common journeys
Most education libraries are a pile of articles, not a route. NEXT reads the questions customers actually ask while ramping — from tickets, calls, chat logs, and onboarding notes — and finds the order successful customers moved through. It hands your education team a proposed learning path: the real sequence of steps, where people stall, and which accounts it would help.
The library isn't the problem. The problem is that a new admin has forty good articles and no idea which one to open first.
What the proposed learning path looks like
Proposed path: first 30 days for mid-market admins
Common ramp sequence
The order most successful accounts moved through, drawn from how they actually onboarded:
Connect the first data source
Invite the team and set roles
Build the first report
Set up alerts
Share a result with a stakeholder
Where customers stall
Between inviting the team and building the first report. Accounts get everyone into the workspace, then stop — they don't know what to make first, so nothing gets shared and the workspace goes quiet.
What customers ask, roughly in order
"I've got everyone invited but I have no idea what I'm actually supposed to build first."
"There are forty help articles and none of them say 'start here.' Which one matters?"
Affected accounts
About 60 accounts ramped through this sequence last quarter. 24 stalled at the same point, before building anything.
Commercial exposure
The stall sits on first-value activation — the step that most predicts whether a new account renews. Roughly $480K ARR onboarded through this path in the period.
Why a path would help
The successful journey is consistent and repeatable, but the library doesn't reflect it. A guided path that bridges "team invited" to "first report shared" would target the exact point where ramp stalls.
Signal strength
Strong and consistent for mid-market admins. Thin for self-serve SMB, where ramp behavior looks different and the sample is smaller.
Example path assembled from grouped onboarding questions and ramp data across similar accounts.
How NEXT does this
NEXT reads where customers ask for help as they ramp — onboarding tickets, support chats, kickoff and check-in calls, and the notes your team already keeps. It builds a continuously updated record of which questions come up, in what order, and which accounts moved from one step to the next without stalling. When the pattern is stable enough, NEXT drafts a proposed learning path — the sequence, the stall points, and the accounts it would help — and routes it to your education team ahead of the quarterly content review. You decide whether the path becomes content. NEXT keeps the underlying picture current, so the next review starts from what customers asked this quarter, not last year's guess.
Why guided paths are hard to build from a flat library today
Most education teams know their content should be sequenced. The hard part is knowing the real sequence. The articles were written one at a time, by topic, as gaps appeared — not in the order a customer needs them.
The data that would tell you the order is scattered. Open your product analytics and you see completion rates, not the questions customers asked to get there. Ask an AI assistant and you get the loudest recent thread, not the path most customers actually followed. Neither comes looking for you when ramp behavior shifts — you have to remember to go check.
And the detail that would tell you what to build decays at every step. The exact question a customer asked between step two and step three gets paraphrased into a ticket, summarized in a QBR note, and half-remembered by the time anyone plans content. What reaches the quarterly review is a vague sense that "onboarding could be better," not a sequence you can build against.
NEXT pushes the path to the people who build training. It doesn't wait for someone to open a report and reconstruct the journey by hand.
How this compares to the tools you already know
Approach | Where the evidence lives | What the education team does at decision time |
|---|---|---|
Flat help library | In the articles themselves, unordered | Guesses the sequence, or leaves customers to guess |
Product or funnel analytics | In dashboards, as completion rates | Sees where ramp drops off, not what customers asked |
AI assistant | Wherever you ask, one query at a time | Gets the loudest recent thread, not the common path |
NEXT | A current record of ramp questions, in order | Reviews a proposed path drawn from real journeys |
What changes for your quarterly content review
Today, the content review starts from a blank agenda. Someone pulls a few tickets, skims a churn note, and proposes new articles based on what they remember hearing. The list is plausible but unanchored — you can't tell which gap actually blocks ramp and which is just a vocal account.
With NEXT, you open the review and the proposed path is already there: the sequence customers moved through, the point where 24 of them stalled, and the two questions they asked at that point in their own words. You're not reconstructing the journey. You're deciding whether to build the bridge between "team invited" and "first report shared," because the data already told you that's where ramp dies.
The quiet workspace stopped looking like low engagement and started looking like a missing step three. That reframes the whole content backlog — you sequence around the stall instead of writing another standalone article nobody opens in order.
The call on what to build, and in what order to ship it, stays with your team. NEXT supplies the journey; the curriculum is still yours.
Downstream effects
CS gets a sharper handoff. When the path names the stall point, onboarding managers can intervene at step three specifically, instead of nudging accounts that are actually fine.
Content priorities get defensible. "Build the first-report walkthrough" carries the 24 stalled accounts and the ARR behind them, so it survives the trade-off against lower-impact requests.
The path stays current between reviews. As ramp behavior shifts — a new feature changes the sequence, say — the next review reflects it without anyone re-pulling the data.
Where the human stays in control
NEXT proposes; it does not publish. A path only reaches your review when the underlying questions are consistent enough to suggest a real sequence, not a coincidence. You set how strong that pattern has to be before a path is drafted, and you can require a person to confirm any path before it's routed.
That's configuration work — deciding what counts as a stable journey and for which segments — not approval work on every individual ticket. You tune the thresholds once and adjust them as your onboarding changes.
What the output depends on
The path is only as good as the ramp signal underneath it. A few things to get right:
Source coverage. NEXT needs to read where customers actually ask during onboarding — support tickets, onboarding calls, and chat. If kickoff calls aren't recorded or notes aren't captured, the early-ramp steps will be thin.
Segment separation. Mid-market admins and self-serve SMB ramp differently. Mixing them produces a path that fits neither. Tell NEXT which segments to treat separately.
Enough volume. A path drawn from six accounts is a hunch. Set a floor so paths are proposed only where the journey repeats.
Review timing. The draft should land before the quarterly review, not during it, so the team has time to weigh it against the rest of the content roadmap.
Where this breaks down
Low onboarding volume
If only a handful of accounts ramp each quarter, there's no common journey to find. NEXT will hold rather than propose a path from too few data points — which is correct, but it means new or niche segments won't generate paths until volume builds.
Silent customers
The method depends on customers asking questions. Accounts that ramp without ever contacting support or appearing on a call leave no trace, so a path built only from vocal accounts can miss how the quiet, successful ones actually moved.
Mixed segments
If two very different customer types are pooled together, the proposed sequence blurs into something that helps neither. This is a setup choice, not a flaw — separate the segments and the paths sharpen.
Sequence drift after a product change
A major UI or feature change can reroute how customers ramp. For a period afterward, the historical path reflects the old journey. NEXT re-learns the new sequence as fresh questions come in, but the team should treat paths as provisional right after a big release.
FAQ
How is this different from a product analytics funnel?
A funnel shows where users drop off in the product — completion rates by step. It doesn't tell you what customers were trying to do or what confused them. NEXT reads the questions customers ask while ramping, so you see not just that they stalled between two steps, but what they said they were stuck on. That's the difference between knowing a number moved and knowing what to build.
Does NEXT decide what content we create?
No. NEXT proposes a path drawn from real journeys and keeps it current. Your education team decides whether the path becomes content, what to build first, and how it fits the rest of the roadmap. The curriculum call stays with the people who own it.
How many accounts does NEXT need before it proposes a path?
You set the floor. NEXT only drafts a path where the journey repeats across enough accounts to be a real pattern rather than a coincidence. For small or new segments, that means no path is proposed until volume builds — which keeps you from sequencing content around a handful of outliers.
What if our onboarding just changed?
A major product or process change can reroute how customers ramp, so a historical path may reflect the old journey for a while. NEXT re-learns the sequence as new questions come in. Right after a big release, treat proposed paths as provisional and let the fresh signal accumulate before committing content to them.
Can it tell mid-market and SMB journeys apart?
Yes, if you set it up that way. You tell NEXT which segments to treat separately, and it builds a distinct path for each. Pooling very different customer types produces a blurred sequence that fits neither, so segment separation is one of the main things to get right before you turn it on.