Generative Brand Imagery
C-Level Keynote Design

The creative concept for DeepL's annual company (CKO2025) event called for imagery that communicated future, growth, and discovery at a scale. I designed a generative workflow to produce all imagery using AI tools, selecting, prompting, and refining outputs until the visual world matched the creative brief.

As a sub-brand context, the CKO visual world allowed more imaginative territory, more evocative, expansive, and rich with possibility. Every image was produced, curated, and integrated into the final keynote and event materials.

Lummi AI → Figma Weave → Figma

✦ SKILLS: SYSTEMS THINKING · KEYNOTE DESIGN · AI WORKFLOWS
✦ TOOLS: FIGMA, FIGMA WEAVE, LUMMI AI
✦ TEAM: ANA M. RODRIGUES (BRANDING LEAD)

Generative Brand Imagery
DeepL Voice for Conversations

DeepL Voice for Conversations required product imagery with very specific visual behaviour: people in real conversation contexts, using earphones and mobile devices, in a way that felt natural, diverse, and on-brand. Stock photography couldn't deliver the precision or consistency required.

The solution was a structured prompt architecture built as a Gemini Gem — a reusable, brand-trained system that generates consistent outputs across multiple variables without starting from scratch each time.

The architecture operates in layers:

Brand layer: visual vocabulary, environment, photographic style
People layer: diversity, body behaviour, natural interaction
Device layer: airpods, mobile, conversation context

Gemini Gem (layered prompt architecture) → Figma Weave → Figma

✦ SKILLS: SYSTEMS THINKING · KEYNOTE DESIGN · AI WORKFLOWS
✦ TOOLS: GEMINI AI, FIGMA, FIGMA WEAVE, LUMMI AI

Go up

Selected work

Next