UX Encyclopedia

Flow Selection Guide

You don't pick a flow type by name — you pick it by the question you need answered. Find your question below, start with that artifact, and expect to add 2–3 companions: the starting artifact is almost never the only one a real project needs.

Choose by question

About the user's path

  • What path does the user take, all routes included? → User flow (User Flow)
  • What's the simplest way to complete this one task? → Task flow (Task Flow)
  • Which screens, in what sequence, with what on them? → Screen flow / wireflow (Screen Flow / Wireflow)
  • How do users move through the product's structure? → Navigation flow (Navigation Flow)
  • How is everything organized and named in the first place? → Information architecture (Information Architecture)

About experience over time

  • What does the user do and feel across a whole scenario? → User journey map (User Journey Map)
  • What's the whole customer relationship — marketing, sales, support, renewal? → Customer journey map (Customer Journey Map)
  • What's the person's real-world experience beyond our touchpoints? → Experience map (Experience Map)
  • Full arc from first ad to renewal in one artifact? → End-to-end flow (End-to-End Flow)

About the system and what can go wrong

  • What happens behind the scenes to deliver this? → Service blueprint (Service Blueprint)
  • Where does logic branch, and on what conditions? → Decision flow (Decision Flow)
  • What states can this screen/component be in? → State flow (State Flow)
  • What can go wrong, and how do users recover? → Error flow (Error Flow) + State flow (14) together
  • What does the user see before there's any data? → Empty-state flow (Empty State Flow)
  • How does information move through the system? → Data flow (Data Flow); how do services/events behave? → System flow (System Flow)

About people and operations

  • How do multiple roles/actors interact on one task? → Multi-user flow (Multi-User Flow) / Swimlane (Swimlane Flow)
  • What are the operational steps of the business process? → Process flow (Process Flow); how roles complete work together → Workflow diagram (Workflow Diagram)
  • Who is allowed to do what, and what do the blocked see? → Permission flow (Permission Flow)
  • When and why do we interrupt the user? → Notification flow (Notification Flows)

Named journeys with their own files

  • Sign-in, passwords, passkeys, 2FA → Authentication flow (15); new-account creation → Registration flow (18); first-run → Onboarding flow (16); payment → Checkout flow (17).
  • Search, filter, forms, upload, export, subscription, cancellation, support, AI agent actions → Specialized flows catalog (32).

New-channel questions

  • How does a dialogue with an AI/bot unfold? → Conversation flow (26).
  • Voice, hands-free, no screen? → Voice interaction flow (27).
  • Movement and action in 3D space (VR/MR/AR)? → XR interaction flow (28).

Then: review any flow against Flow Design Checklists before wireframes, and steal shapes from Examples Library.

Worked scenarios: picking the combo

New SaaS feature (e.g., "shared project templates")

  1. Task flow (03) — the ideal path for the one new job, to keep scope honest. 2. User flow (02) — all entry points and routes, including discovery of the feature. 3. State flow (14) + empty-state flow (20) — templates have a hard blank-start problem. 4. Permission flow (22) — "shared" means roles; who can create, edit, delete a template? Skip journey maps: the relationship isn't changing, one feature is.

E-commerce checkout redesign

  1. Checkout flow (17) — the specialized route itself, guest path included. 2. Error flow (19) — declines, timeouts, address validation failures are where checkouts die. 3. Service blueprint (07) — payment processor, fraud check, inventory, fulfillment must line up with each screen. 4. State flow (14) — every step's loading/ error/success behavior. A journey map (04) helps only if the problem is upstream of checkout (anxiety, trust) rather than in it.

AI assistant feature (agent that acts for the user)

  1. Conversation flow (26) — intents, slots, clarification, escalation.
  2. AI agent flow (32 § AI Agent Flow) — plan preview, progress, consequential-action gates, undo. 3. Error flow (19) — wrong answers, refusals, and misfires are the normal case to design, not the exception. 4. Data flow (24) — what user data reaches the model, what's retained, what trains it (the privacy disclosure depends on knowing this). Add voice interaction flow (27) only if it ships on a spoken surface.

One rule across all three: the artifact you start with frames the problem; the companions catch what it can't see. Budget for the set, not the single diagram.

↑↓ to navigate · Enter to open · Esc to close