01 / QUESTION
Can AI extend the shoot without losing the brand?
The August 2026 test explored whether Figma Weave could extend brand photography into new environments and compositions. The question went beyond an impressive individual image: could a workflow preserve talent likeness, deliver consistent quality, and produce assets the creative team could use?
02 / METHODS
Compare workflows, not just outputs.
Three approaches were tested: avatar and character sheets, background replacement, and a custom hybrid workflow. Each was evaluated for creative output and operational readiness, including likeness, realism, integration, repeatability, composition usability, and retouch burden.
Three approaches. A shared set of criteria.
01 / AVATAR
Face + full-body photography → Character sheets → Generated brand environments
02 / BACKGROUND REPLACEMENT
Posed photography → Isolated talent → New environment
03 / CUSTOM / HYBRID
Existing brand photography → Controlled edits → Extended brand compositions
03 / AVATAR
Strong moments. Inconsistent control.
Character sheets built from face and full-body photography produced some strong isolated images. Across generations, likeness could drift and expressions or environments could feel generic. The recommendation was to continue R&D rather than define a production recipe.
04 / BACKGROUNDS
Preserve the person. Test the setting.
Isolating photographed talent and placing them in new environments gave more control over expression and likeness. Fixed poses limited flexibility, while lighting, scale, perspective, and contact points remained integration challenges. The approach showed potential for selected post-production uses.
05 / HYBRID
Start with the brand’s strongest inputs.
The custom workflow extended existing brand photography through controlled changes to wardrobe, talent, environments, props, poses, and compositions. It produced the most consistent and ownable results in the test, with stronger likeness retention and repeatability.
06 / FINDINGS
Advance the strongest workflow, with a clear next test.
The hybrid approach earned the highest internal overall rating: 4.2 out of 5. It was recommended for the next production test. Stills showed promise, while motion, broader asset scaling, and final production polish remained unproven. Further model testing, workflow documentation, and usage guidelines were proposed next steps.
Internal test scorecard
Comparable models could produce closer results in Runway. The test favored Weave’s workflow and Figma integration; it did not establish a universal model-quality winner.
