Role: Design, Art Direction
Client: DoorDash
DashPass AI Images
Made at DoorDash (Superette)
Head of Design Adriel Teles
Creative Lead Ruby Lee, Chris Ballard
Designer/Art Director Chris Ruppelt
Producer Jeremy Lewis
Partner Agency Tool North America
Our DashPass team faced a critical need for generic imagery and videos relevant to DashPass partners. Brand partners often needed very contextual stock and the costs could quickly add up.
Rather than resolving these content requests on an a la carte basis, we partnered with Tool to generate a suite of images. They were generic enough to be flexible and contextual enough to meet the specific needs of certain partners.
Approach
The approach was similar to a traditional shoot; but our notes on action, talent blocking, wardrobe, lighting, texture, and tone would instead be used as data points for creating tailored prompts.
01 Shot Objective
Input a set of data points into a shot list matrix that acts as a brief per image.
02 Composition & Framing
Focus on talent blocking, food, propping, and styling. Provide feedback to regenerate based on feedback.
03 Fine Tune
Details. Regenerate if necessary, but primarily clean up artifacts, check consistency, etc.
04 Delivery
Final generative images are upscaled. Light and color are retouched. Images delivered to DoorDash.
Groceries & Food
We needed to show people and families receiving and unpacking groceries as well as more evergreen food deliveries.
Travel & Lifestyle
These images were integral for partners like Lyft. We wanted to illustrate uses cases where a consumer might be using both services in various scenarios — an order on the way home from the airport, restocking the fridge, etc.
Streaming & UI
Images that combined the DoorDash UI and different devices for streaming content needed to be templatized so we could drop in logos, key art, or UI mockups.
Cost Comparison
At a lower funnel level we found that were significant savings between a traditional and generative shoot. For surfaces like landing pages, emails, etc the output was useful and contextually specific for our partners.
Traditional Shoot:
Budget Estimate $300,000
Shoot Days 2
Talent Limited Cast (6 max)
Locations Limited (2-4 max)
Deliverables/Output 45–60 Total
Usage/Licensing 1x Year, No Broadcast, Print, or OOH; Limited pass-through-rights
Estimated Cost per Asset $5,000 - $6,667
Generative AI Production:
Budget Estimate $140,000
Shoot Days NA
Talent No Firm Limit
Locations No Firm Limit
Deliverables/Output 110 Total
Usage/Licensing 2x years, No Broadcast or OOH; Unlimited pass-through-rights
Estimated Cost per Asset $1,272