Industry
Healthcare AI
Company
Infinitus
Modular AI becomes visual
A metaphor for a multimodal, multi-model platform
Infinitus was already a leader in voice AI for healthcare, but as competitors began to emerge, we needed a brand system that connected with what made this platform fundamentally different.
Not a simple ChatGPT wrapper, Infinitus was built on a multimodal, multi-model system — with layered guardrails including a discrete action space and a knowledge graph derived from millions of real calls. Its selective, modular nature was its superpower. The brand system needed to reflect that: dynamic, adaptive, and engineered.
That led to the brand cube system — glowing 3D forms that could shift, recombine, and signal infinite configurations. A metaphor for a platform built from interoperable parts, where each component has a role. The system gave a nod to existing brand elements (like isometric illustrations and the square identity) while evolving it for a more expressive, dimensional future.
We began with AI-generated explorations, refined in Blender, and rolled the visual system into motion, demo videos, and beyond.
A visual system that scales with the needs of the marketing and sales teams
Beyond the metaphor, the system served a practical role: pairing technical content with an abstract, mood-setting aesthetic to avoid overwhelming the viewer with dry detail. It offered a clean visual companion to product demos — call recordings — without competing for attention.
The graphic system also improved execution speed. With a defined set of flexible visuals, the marketing team could create on-brand assets quickly, from one-pagers to social images. It enabled scalable templates and made it easier for non-designers to stay consistent across communications.
Because the visuals weren’t tied to a single feature or use case, they could support any message — giving the brand cohesion without limiting creativity.
2x
Assets produced by the marketing team as a result of the new design system in one year
Credits:
Sarah Baicker, Content
