Press play on a document.
A review tells you what a reader thinks of your doc. A run tells you whether someone can actually follow it. Subjourney builds the journey your docs describe, then sends simulated personas through it to find out.
01Import your sources
Start with what you already have.
Drop in the specs, PRDs, policies, process docs and diagrams your team already works from. No templates, no rewriting into a new format. Subjourney reads them as they are.
02Model the journey
Your docs describe a journey. We draw it.
Every doc implies steps: someone signs up, requests approval, hits an edge case, gets an answer. Subjourney turns those implied steps into a journey graph, with the paths, decisions and outcomes your docs promise.
It also builds the personas who'd walk it, each with their own goal, context and constraints. While it builds, it flags the obvious holes: steps with no exit, outcomes nothing leads to.
Open any paragraph and see which steps it shapes. Open any step and see which paragraphs it came from.
03Run it
Watch a persona try to get where they're going.
Pick a persona and press run. They move through the graph node by node, reading only what your docs give them, making the choices a person in their position would make.
When the docs give them what they need, they move on. When they don't, they stall, and you see it happen.
04Fix at the source
Every stall has a paragraph.
Findings aren't a vague score or a list of suggestions. Each one names the persona, the step, and the paragraph that caused the stall, or the paragraph that should exist and doesn't.
- MissingThe doc never covers this case.
- AmbiguousThe doc covers it two ways, or not clearly enough to act on.
- ContradictoryTwo docs disagree, and the persona can't satisfy both.
- Dead endA path that leads nowhere near the promised outcome.
05Re-run to verify
Close the loop, then prove it's closed.
Send findings straight to the agent you build with, or hand them to whoever owns the doc. When the fix lands, run again. Same persona, same goal. If they make it through, you know.
This is the part code has had for decades: change something, run the tests, see green. Subjourney brings it to the documents your organisation runs on.
Meet who walks your docs.
Run the PRD before engineering builds it.Product teams
Solo userSubscriber
Switching to annual billing.
- Motivation
- Lock in the annual discount now the tool is part of every week.
- Frustration
- Being charged twice when a plan changes mid-cycle.
- Need
- A plain note on what happens to the month already paid.
- Age29
- Income$58,000
- EducationBachelors Degree
- EmploymentEarly Career
- Communication styleCasual
- Social engagementActive Contributor
- Risk appetiteRisk-Tolerant
- Eco awarenessEco-Conscious
- Work-life balanceFlexible
- Brand loyaltyActive Loyalty
- Price sensitivityValue Seeker
- Tech adoptionEarly Adopter
- Tech literacyAdvanced
Questions
Aren't these just AI opinions about my doc?
No. A persona doesn't review your doc. It tries to use it, one step at a time, and only knows what the doc tells it. You get evidence of where it broke, not feedback on how it reads.
Does this replace user research?
No. Simulated personas catch what's broken in the doc before real people ever see it, so your research time goes to the questions only people can answer.
What if my doc isn't finished?
Run it anyway. Early runs on a rough draft are where the most useful stalls show up.
What kinds of docs work?
Anything that describes how someone gets from a starting point to an outcome: PRDs, service blueprints, policies, SOPs, onboarding guides, support playbooks, agent instructions.
Bring one doc. See where it breaks.
Pick the doc you're least sure about and run it.