How KINESSO helped Amazon Fresh cut production time by 93% using Adobe Firefly.

Aaron Mitchell Finegold

07-16-2025

In a global retail market shaped by seasonal shifts, localized consumer preferences, and fast-changing commerce trends that demand continuous adaptation, Amazon Fresh is redefining how grocery brands scale both content and creativity to stay connected with customers around the world.

An Adobe survey* found that nearly two-thirds of marketers expect content demand to surge fivefold by 2026, forcing creative teams to accelerate production to keep up with increasing volume and velocity. For global brands like Amazon Fresh, that expectation is already a reality — making it critical to build smarter systems that can adapt and scale content across formats, channels, and regions.

Amazon Fresh campaigns evolve constantly, with assets needed across global markets and a wide range of formats like app banners, social media, and in-store displays. The creative team was challenged to keep pace, as traditional production methods alone couldn’t deliver the volume or velocity required to maintain high standards and consistency across regions and platforms.

To meet the growing demand for digital assets, Amazon Fresh partnered with KINESSO, a performance-first agency within IPG Mediabrands, to bring new speed and flexibility into its creative workflow. Together, they implemented Adobe Firefly Custom Models to train generative AI to produce on-brand visuals that plugged seamlessly into campaigns and cut production time by 93%.

Bringing new creative worlds to life on established foundations.

The Amazon Fresh marketing team regularly develops seasonal and evergreen campaigns based on beautifully staged brand photography. As content demands increased, with more channels, tighter timelines, and the need to personalize creative for different regions and promotions, the process became harder to scale.

Over time, marketers and creatives began spending more time managing logistics than testing new ideas, and the pace of production had begun to bottleneck. They needed a way to build on existing evergreen assets and deliver visuals faster, with more variation and without compromising brand consistency.

KINESSO explored various ways to help Amazon Fresh deliver more campaign content without increasing production requirements. One of those was to embrace generative AI through Adobe Firefly — a commercially safe family of text-to-image models.

By training the model with approved Amazon Fresh assets and creating a Custom Model, the team was able to generate entirely new visuals that looked and felt like the original campaigns. This approach turned lead times from weeks to days. These weren’t generic AI outputs. They were distinctive, brand-aligned assets that seamlessly extended the visual system and brand design language Amazon Fresh had already built.

Many of the visuals drew from seasonal themes, such as holiday tables or school lunch spreads, and they reflected the brand’s playful, fresh aesthetic — even when the ideas were unexpected.

“Where else other than in Adobe Firefly can you create a Santa hat made of strawberries or a snowman made of cheese, and have it feel totally on brand?” says Andy Duke, group director of creative and experience design at KINESSO.

Enhancing creative flexibility and shifting focus from production to possibility.

With Firefly Custom Models, Amazon Fresh could quickly adapt creative concepts across different events and audiences. Teams could explore new iterations using simple prompts while maintaining high levels of brand fidelity and production quality.

93 percent faster turnaround time on 262 unique images than in previous production processes

The composition reference image feature, which allows designers to guide the overall layout of a generated image, gave the team even greater control, guiding layouts and protecting space for critical campaign elements like copy, CTAs, and logos. But the real breakthrough was creative freedom.

By integrating Firefly Custom Models into their workflow, the Amazon Fresh team shifted from rigid planning cycles to a more flexible, test-and-learn approach. Designers could quickly generate concepts, explore variations, and try unexpected ideas. This approach increased efficiency and speed dramatically.

For example, in the Christmas campaign, the team produced 262 unique images with a 93% faster turnaround time than in previous processes. The brand gained room to play — and with it, the ability to adapt campaigns mid-flight while maintaining consistency across touchpoints.

“We’ve found ways to scale without compromise. Firefly lets us explore, iterate, and refine — all while keeping our creative standards sky-high.”

Andy Duke

Group Director of Creative and Experience Design, KINESSO

Unlocking a scalable brand storytelling strategy.

The transformation resulted in faster turnarounds, standardized visuals, and a more nimble pipeline. For KINESSO, it validated a scalable model for creative operations — one that puts performance, precision, and imagination on equal footing.

The partnership between Amazon Fresh and KINESSO reflects a shift in how leading brands are approaching content creation — not by replacing traditional workflows, but by enhancing them. With Firefly Custom Models, teams gain the flexibility to move faster, iterate more freely, and maintain brand consistency at scale in an increasingly dynamic marketing landscape.

Explore how Adobe Firefly Custom Models help brands like Amazon Fresh scale content with speed, precision, and consistency.

Watch the Adobe Firefly webinar to see how leading brands are transforming creative operations.

*Adobe Survey of 2,841 marketers in markets including the U.S., Australia, France, Germany, India, Japan, and the U.K. (fielded from Feb. 27 to Mar. 7, 2024)

https://business.adobe.com/fragments/resources/cards/thank-you-collections/firefly