The snapshot
1
DREST faced content creation bottleneck with manual process taking half a day per designer across 300+ luxury brand partners
2
Through AWS-funded ADAPT program, Devoteam validated AI feasibility and built prototyping tool in five-week engagement
3
‘Dresty’ platform now gives DREST’s team self-service access to test AI models, establishing foundation for future model exploration
About DREST
DREST is a luxury fashion styling mobile app that democratises access to high-end fashion through AI-powered innovation. Operating as a multi-brand environment featuring 300+ prestigious luxury brands, the platform connects users with iconic labels, emerging designers, and renowned hair and makeup artists. Recognised by Fast Company as one of the World’s Most Innovative Companies 2021, DREST transforms what was once exclusive to industry insiders into an accessible digital experience for fashion lovers globally.
The Challenge
Behind DREST’s user experience lies a complex content creation workflow. DREST’s graphic designers manually edit clothing assets onto model images using Photoshop-like tools, adjusting product images from 300+ luxury brand partners to fit model avatars. This meticulous process takes approximately half a day per person, with five full-time designers dedicated to the task.
While ensuring exceptional visual quality that luxury brand partners expect, this manual approach presents significant scalability barriers. The process is technically complex, requiring approximately six months to train new designers. Additional challenges include layering garments, applying realistic drop shadows, and maintaining high visual accuracy across diverse clothing types.
The Solution
Devoteam developed a comprehensive AI-driven solution through our ADAPT (AI & Data Acceleration Program for Transformation) framework, in collaboration with and powered by the AWS Generative AI Innovation Center. This exclusive program leveraged AWS funding available through our Strategic Collaboration Agreement (SCA) in addition to dedicated Proof of Concept (PoC) funding for discovery and prototyping, to validate AI feasibility while ensuring alignment with DREST’s creative requirements.
The project defined specific success criteria: maintain quality standards, achieve significant processing time reduction, successfully fit diverse clothing items, and validate results against manual processes.
Structured Discovery & Multi-Model Development
The solution evaluated multiple AI models to determine optimal approaches for different scenarios. This evaluation led to a structured two-project approach:
- Project 1 focused on proving whether AI could fit clothing to models, prioritising quality of fitting and texture/pattern fidelity.
- Project 2 involved building “Dresty,” a prototyping tool that provides DREST’s content team with self-service access to test AI and pre-trained ML models. This approach helps with training and allows them to scale to other models in future without investing in expensive model training.
AWS Cloud Infrastructure
Built on AWS Fargate ECS with Lambda functions, S3, RDS Aurora, DynamoDB, and enterprise-grade security through Secrets Manager, WAF, KMS, and CloudWatch monitoring.
The Benefits
The ADAPT methodology delivered measurable results validating both technical feasibility and business value:
Technical Validation
- Processing Speed: Image generation demonstrated in approximately 1-minute timeframes
- Image Quality Results: Demonstrated excellent texture, quality and realism in AI-generated images
- ‘Dresty’ Platform: Created reusable framework for ongoing AI model evaluation with batch processing capabilities
- Scalability Foundation: AWS infrastructure established supporting potential future implementations
Strategic Business Impact
- Innovation Leadership: Positioned DREST at forefront of AI adoption in digital fashion
- Growth Enablement: Created scalable technical foundation supporting 300+ brand partnerships
- Future-Ready Architecture: Comprehensive infrastructure capable of supporting advanced AI implementations
- Strategic Investment Model: leveraged AWS Strategic Collaboration Agreement (SCA) and Proof of Concept (PoC) funding through ADAPT’s collaborative framework, combining AWS infrastructure with Devoteam’s AI expertise to reduce financial barriers while enabling upfront feasibility validation and accelerated development
Complete proof-of-concept validation delivered with technical documentation, architectural handover, and 6-month technology review planned.
This is a project Drest had had challenges with in the past due to the quality of the output. Devoteam was given a hard brief to develop a PoC that would achieve our desired quality threshold. They did an excellent job, not only hitting the quality but also delivering a tool that allows us to test and evaluate new models as they become available. The team at Devoteam were great to work with, they kept us up to date on progress and delivered on time and to budget. Overall an excellent experience.
Chris Wright
Chief Technology Officer at DREST
Why Devoteam?
As a Premier AWS Partner with Generative AI competency, Devoteam’s ADAPT program provided strategic advantages:
Proven Framework
ADAPT’s work-backwards methodology from business vision to implementation, with upfront ROI validation and funded discovery approach through AWS Generative AI Innovation Center partnership.
Technical Excellence
Deep AI expertise across multiple ML models and frameworks, combined with creative industry understanding that solutions must enhance rather than replace human creativity.
Strategic Investment
AWS-Devoteam collaboration combines infrastructure funding with AI expertise to provide cost-effective, evidence-based validation of transformation initiatives.
Ready to transform your creative workflows with AI?
Discover how Devoteam can help your organisation leverage cutting-edge AI technologies to enhance creativity, improve efficiency, and maintain exceptional quality standards.