Business Problem

• Inefficient Marketing Content Creation: The traditional process of designing marketing visuals was time-consuming and resource-intensive.
• Lack of Personalization: Generic marketing materials did not align with the diverse needs of Mahindra’s product segments.
• Scalability Issues: Managing large-scale marketing campaigns with high-quality visuals required automation.

About Mahindra

Mahindra is a global conglomerate with diverse business interests, including automotive, farm equipment, IT services, and financial services. The company is committed to innovation and digital transformation, leveraging AI-driven solutions to enhance operational efficiency and customer engagement.

Solution

AI-Powered Marketing Content Creation – Enabled Mahindra to generate high-quality marketing visuals and text using AI, reducing dependency on manual design efforts. Tech Stack: AWS Bedrock (Claude 3.5 Sonnet for text generation, Stable Image Ultra for image generation), React.js for frontend development.

On-Demand Image Regeneration – Enabled Mahindra’s marketing teams to regenerate AI-generated images dynamically, providing variations for different campaign needs. Tech Stack: Stable Image Ultra API, Dockerized backend for scalable processing, MongoDB for tracking generated assets.

LLM used: The LLMs used in this case study were Claude 3.5 Sonnet for AI-driven text generation and Stable Image Ultra for high-quality image creation, both integrated via AWS Bedrock to streamline Mahindra’s marketing workflows.

Automated Brand Asset Management – Provided a centralized repository for Mahindra’s brand assets, allowing seamless integration of logos, templates, and design elements into marketing materials. Tech Stack: AWS S3 for secure storage, MongoDB for metadata management and retrieval, React.js for UI-based asset management.

Seamless Deployment & CI/CD Integration – Implemented an automated deployment pipeline to ensure efficient updates, scalability, and minimal downtime for Mahindra’s marketing platform. Tech Stack: AWS CodeDeploy, Docker, EC2 (m6i.large) for hosting, Git for version control and deployment automation.

Customizable Text & Background Removal – Allowed users to generate text overlays with the option to remove backgrounds, ensuring flexibility in design customization. Tech Stack: Python-based image processing, AWS Lambda for real-time execution, Stable Image Ultra for background manipulation.

Multi-Aspect Ratio Image Generation – Allowed Mahindra to generate marketing visuals in different aspect ratios, ensuring content adaptability for social media, websites, and print. Tech Stack: Python-based image processing, Stable Image Ultra API, React.js for UI-based selection of aspect ratios.

Architecture

• Frontend
  React.js – Provides an intuitive and responsive UI.
  CloudFront – Handles content delivery for faster access.
  S3 Bucket – Stores frontend assets and static content.
• Backend
  EC2 Instance (m6i.large - ap-south-1) – Hosts backend services.
  Python 3.10+ – Core programming language for backend logic and APIs.
  MongoDB – Stores project details, templates, and generated images.
• AI Models (via AWS Bedrock)
  Claude 3.5 Sonnet (ap-south-1) – Used for AI-driven text generation.
  Stable Image Ultra (us-west-2) – Generates high-quality marketing visuals.
• Storage
  S3 Bucket (ap-south-1) – Stores generated images and other assets.
• Version Control
  Git – Manages source code for both frontend and backend.