Question 1

What makes a good user story?

Accepted Answer

A good user story names a specific persona (not 'user'), describes a verifiable action, and ends with a measurable outcome that delivers real user value. It should pass the INVEST test — Independent, Negotiable, Valuable, Estimable, Small, Testable. Vague verbs like 'manage' or 'use' are red flags. Prefer concrete verbs (save, share, cancel, export) and name the artifact the user acts on.

Question 2

What does "As a / I want / So that" actually mean?

Accepted Answer

It's the Connextra template, coined at Connextra Ltd in the early 2000s and popularised by agile teams because it forces three things into one sentence: the who (persona), the what (action), and the why (outcome). The 'so that' clause is the one most teams skip — and the one that saves the most time, because it forces you to justify the feature before building it. If you can't finish the 'so that' clause, the story probably isn't ready.

Question 3

How many user stories should a feature have?

Accepted Answer

Usually 3–7 for a discrete feature. Fewer than three often means you're describing an epic or hiding complexity inside vague verbs. More than seven usually means you're mixing concerns — try splitting by persona, by path (happy vs edge), or by release. The INVEST principle's 'Small' hint: a story you can't ship in a single sprint should probably split.

Question 4

What are T-shirt size estimates?

Accepted Answer

Relative estimates — XS, S, M, L — that sidestep the precision trap of hours or story points. A T-shirt size is a quick judgment of size relative to other work, not a commitment. They beat hour estimates because they're faster to agree on, more honest about uncertainty, and don't pretend the team can predict individual effort down to the hour. Convert to throughput over time instead.

Question 5

How is this different from asking ChatGPT for user stories?

Accepted Answer

This generator returns a structured JSON schema — persona, action, outcome, priority, estimate, rationale — every time. ChatGPT will give you prose that changes format from query to query. We also prime the model specifically for MoSCoW prioritisation and INVEST-shaped output, and the stories are linked to follow-up tools (acceptance criteria, personas) so the output keeps compounding.

Question 6

Can I use these stories in Jira or Linear?

Accepted Answer

Yes — copy the card content directly into any backlog tool. Jira, Linear, Asana, and Productboard all accept the Connextra format out of the box. If you capture your email we'll send a Markdown + CSV export ready for bulk import. The full VibeMap app also exports directly to Linear and Jira if you need a tighter loop.

Question 7

Do you store my input?

Accepted Answer

Your input is retained for up to 30 days purely for abuse monitoring and generation quality review. We never train models on it and never share it with third parties. You can request deletion any time by emailing support@vibemap.ai. If you capture your email we'll include an unsubscribe link with every message.

User Story GeneratorTurn a feature idea into testable user stories in seconds

Acceptance Criteria Generator

Persona Generator