IDEATION: FROM HMW TO IA
I translated my research into focused How Might We prompts around first-minute capture, trust, and expressive input, then ran timed Crazy 8s to generate breadth and clustered patterns to converge. To keep concepts accountable to the half-awake context, I set measurable guardrails: ≤5s to first input and <2 required decisions before capture, grounded in Hick’s Law and progressive disclosure to minimize cognitive load. I tested my sketches with the personas, chose what worked best, and built a lean IA, a straightforward notifications model, and three priority flows (Morning Prompt, Export to Therapy, Expressive Capture) to wireframe.
Bianca: From Problem to Goal
From interviews and empathy maps, I framed Bianca’s problem as “I wake groggy and forget fast; most apps ask too much, too soon.” I translated that into a goal: enable first-minute, one-tap capture that works in low light and low cognition-voice-first, autosave with save-as-draft on exit, and mood at entry for quick structure.
Sonia: From Problem to Goal
Sonia’s problem centered on trust: “I need to share a dream for therapy without ambiguity about what’s sent or how private it is.” The goal became a clear, intentional export: choose what to include, add a brief therapist note, Mark as Sensitive when needed, and receive an explicit sent confirmation with a visible trail. The outcome is confident, low-anxiety sharing that supports therapy without adding cognitive load.
Alex: From Problem to Goal
Alex’s problem was fidelity and flow: “Rigid templates and mode switches break my expression; fragments lose texture.” I set a goal for expressive, multimodal capture-voice + quick sketch/audio available inline, mood chips at entry, and persistent autosave so ideas evolve without interruption. The result is uninterrupted capture that preserves creative nuance and stays searchable for later exploration.
ALEX - HMW (HOW MIGHT WE) FOR EXPRESSIVE CAPTURE
For Alex, I asked: How might we let him capture ideas in different ways without breaking his flow? I explored sketch while recording, inline audio, and mood-at-entry. I scored each against my guardrails (under 5s to first input, fewer than 2 choices), Alex’s need for creative texture. The winner: voice-first capture with autosave, a quick inline sketch pad, and mood chips at entry, so he can express more without pop-ups or setup screens.
SONIA - CRAZY 8’S FOR THERAPY EXPORT
For Sonia, I ran Crazy 8s to reimagine sharing for therapy: select exactly what to include, add a short note, see privacy up front, and receive an explicit confirmation with a visible trail. After eight rapid frames, I converged on a single export sheet with content checkboxes, an inline “Add therapist note”, a clear Sensitive state, and a sent confirmation. I chose this path because it removes ambiguity, keeps cognitive load low, and aligns with Soraya’s outcome-oriented goal while preserving trust.
BIANCA - CRAZY 8’S FOR FIRST MINUTE CAPTURE
For Bianca, the Crazy 8s sprint targeted the half-awake moment: a timed morning prompt, one-tap voice start, large tap targets, save-as-draft on exit, and a gentle fallback for “I forgot.” I selected the concepts that consistently met the guardrails (≤5s, <2 decisions), reduced hesitation, and matched the research on recall decay. The result is a prompt: one tap - recording pattern with warm microcopy and autosave - fast enough to beat memory fade without sacrificing reassurance.
INFORMATION ARCHITECTURE - BACKBONE OF FLOWS AND WIREFRAMES
I translated the personas’ goals into a lean IA that made the first-minute path obvious and everything else predictable. I defined a small set of core objects: Dream, Mood, Attachment, Export, and a simplified settings model (a single notifications toggle with an optional Weekly Summary), then mapped clear entry points: Morning Prompt - Capture, Library - Dream Detail, and Detail - Export.
I defined three clear states: Draft, Sensitive, Exported, so flows could use them without extra choices. That became the blueprint: start in one tap, add mood/attachments inline, and confirm export with a clear activity trail. Locking the hierarchy early (capture - review - share), plus consistent modals (“Save as draft”, “Export”) and tap targets, sped up wireframing. Result: faster iteration, fewer rewrites, and testable wires that stayed within the guardrails.
