Workflow
An AI workflow is a multi-step process that combines prompts, tools, human review, and structured handoffs to reach an outcome, not a single chat reply.
Workflows are how teams ship repeatable design research, content, or launch processes with AI without reinventing steps each time.
Share
What it means
Defined stages (inputs → generate → review → export) often spanning people, systems, and multiple model calls with clear deliverables.
Why designers should care
Workflow UX needs progress, ownership, saved state, and templates. Chat alone rarely carries a team from kickoff to shipped artifact.
Example
A research sprint workflow moves from plan → interview guide → synthesis → readout, with checklist status and links to the right prompt at each stage.
Common mistakes
- • Forcing long processes into one chat thread with no milestones.
- • No save/resume when inference or human review pauses overnight.
- • Workflows that differ by team but share one confusing generic UI.