Lucas Coutinho
Voltar ao Início

GenGuide

AI-powered corporate learning platform for global product teams.

GenGuide Product Design Case Study
PapelSenior Product Designer
Duração6 months (Q3-Q4 2025)
PlataformaResponsive Web Application (Desktop & Tablet)
Equipe1 Product Designer (Lucas), 1 Product Manager, 4 Engineers, 1 AI Specialist
01 / OVERVIEW

Architecture and editorial interface design for a generative AI enterprise solution that converts complex internal documentation and training into personalized, interactive knowledge paths.

Led Product Design end-to-end: strategic alignment, user research, AI interaction guidelines, usability testing, and production design system tokens.

02 / CONTEXT
Context

Global enterprises struggled with low engagement and high drop-off rates in traditional corporate training. Training materials were verbose, static, and difficult to reference during live workflows.

The Problem

Employees lost an average of 4.5 hours per week searching through fragmented docs, while L&D teams spent months producing courses that became outdated quickly.

03 / DISCOVERY & RESEARCH

In-depth research with 28 professionals across 4 regions revealed that the core issue was not a lack of content, but information overload. Users needed contextual, actionable answers with source citations.

04 / KEY DESIGN DECISIONS
DECISION 01Assistive Copilot vs. Fully Autonomous Agent

Determining how AI recommendations should be presented.

Razão:

Prioritized user control and source attribution to mitigate data hallucination concerns and build professional trust.

Trade-off:

Required slightly more explicit user confirmations, but guaranteed transparency over source documentation.

Resultado:

92% increase in user trust scores regarding AI assistant outputs.

DECISION 02Modular Dynamic Card Trails

Structuring learning modules within the core workspace.

Razão:

Replaced deep tree structures with modular, progress-aware content cards.

Trade-off:

Demanded rigorous grid discipline and font hierarchy to maintain low visual noise.

Resultado:

40% reduction in drop-off during the initial 10 minutes of learning sessions.

05 / THE SOLUTION

An editorial, calm workspace combining semantic search, AI executive summaries, and a persistent assistant dock.

  • Contextual assistant dock accessible via keyboard shortcuts
  • Instant mind mapping and chapter-level executive highlights
  • Restrained typographic hierarchy optimized for cognitive focus
GenGuide core workspace showing 12-column editorial grid and assistant dock

Overview of the GenGuide workspace highlighting grid structure and persistent assistant integration.

Visual knowledge mapping exploration view
Immersive reading mode with assistant docked

Left: Visual discovery exploration. Right: Focused reading view with full WCAG 2.2 AA compliance.

06 / SYSTEM & TOKENS

Design System & AI UI Tokens

Built a resilient UI component system with explicit AI states (semantic loading, streaming response, citation tags, and confidence indicators).

Token-integrated AI feedback statesSerif/Sans typography pairing optimized for long-form readingStrict WCAG 2.2 AA contrast standards
07 / OUTCOMES
01

84% voluntary adoption across product and operations departments within 60 days.

02

Significant reduction in onboarding time for new team members.

03

Extremely positive feedback regarding typographic clarity and zero UI clutter.

08 / KEY LEARNINGS
  • AI interactions must emphasize source attribution and human oversight over blind automation.
  • Visual restraint and clear hierarchy dramatically improve retention for dense technical domain data.
  • Continuous cross-disciplinary engineering syncs early in the wireframe phase ensure seamless design token implementation.