L3 describes a system in which a floor plan supplies a face outline and architectural symbols become portrait elements, with related typography.[1] This links the studio’s visual identity to its design work, rather than attaching an unrelated decorative style. That connection is our reading of the design, not a measured customer response.
The agency describes editable avatar software, its use in client documents and social content, and folded packaging for material samples.[1] The D&AD archive credits L3 and Jiazhi in its 2024 entry and describes the use of AutoCAD symbols and an algorithm.[2] These sources document the proposed system; we have not tested the software or verified how widely clients used it.
The agency claims better conversion but provides no comparable before-and-after dataset on the reviewed page. Its Chinese and English follower counts also differ.[1] We omit those numbers and do not treat an award entry as proof of sales uplift. The two pages are not two independent studies of customer outcomes.
Analysis & application
A generative identity becomes useful when teams know what may change and what must remain stable. For a studio, that could mean testing whether a portrait remains recognizable on a small document label, while preserving a consistent symbol vocabulary. Those are design hypotheses to test, not documented rules of Jiazhi’s implementation.
Analysis & application
- Identify a visual element rooted in the service rather than a generic trend.
- Define fixed rules and permitted variations before generating many examples.
- Test real document, small-screen and physical-sample sizes for legibility.
- Let staff create and correct examples; record time, errors and unwanted representations.
- Compare recognition and task completion against the existing identity before making conversion claims. This is our proposed evaluation, not a reported Jiazhi experiment.
FOLLOW THE EVIDENCE
The sources
- 家芝studio L3 Branding · l3branding.com
- Jiazhi Studio — D&AD Awards archive D&AD · dandad.org