Dropt powers white-label fan loyalty and membership platforms for sports teams, athletes, and content creators. Customers including Formula One teams such as Alpine F1 and Haas F1 use the platform to run fan engagement experiences where users earn XP points by reading articles, watching videos, completing quizzes, answering polls, and participating in weekly Pick'em games.
Problem: Manual workflows limited image management system efficiency
For every race weekend, Dropt received bulk image collections from client media teams. Content managers manually reviewed, renamed, categorized, and matched each image to quiz questions, polls, articles, and other fan engagement activities. The process relied on inconsistent file names and manual review, making it difficult to locate relevant images quickly.
Weekly activations often required sorting hundreds of images before selecting a handful for publishing. The existing workflow also lacked semantic search and intelligent recommendations. Content creators could not search using natural language or receive automated image suggestions while creating new campaigns. As media libraries continued to grow, manual image management slowed campaign production and increased operational effort.
Solution: AI powered image management system for sports media
GoML built Dropt using its AI Data Analytics blueprint to deliver AI powered image management system for sports organizations. The platform automates image understanding, metadata generation, semantic search, and intelligent recommendations through Amazon Rekognition, Claude 4.5 Sonnet on Amazon Bedrock, and vector search technology.
The solution processes bulk image uploads, generates rich metadata, creates searchable vector embeddings, and recommends the most relevant images directly within Dropt's CMS through REST APIs.
AI powered image understanding
Example query
"Show images of Pierre Gasly celebrating after the Singapore Grand Prix."
Key capabilities
• AI powered image recognition
• Automated metadata generation
• Intelligent image management system
• Natural language image search
• Vector similarity search
• Smart image recommendations
• Bulk ZIP file processing
• REST APIs for CMS integration
The platform automatically analyzes uploaded images and extracts meaningful information including drivers, teams, race events, locations, objects, activities, and visual context.
Generated metadata includes:
• Driver names
• Teams
• Race circuits
• Season
• Practice sessions
• Qualifying
• Race day
• Podium celebrations
• Team events
• Searchable tags
Intelligent image management system
The image management system understands Formula One media and automatically organizes images into structured categories for faster retrieval.
The platform supports:
• Race weekends
• Testing sessions
• FP1, FP2, FP3
• Qualifying
• Race results
• Driver portraits
• Team activities
• Client-specific categorization rules
The platform automatically:
• Categorizes uploaded images
• Generates searchable metadata
• Creates vector embeddings
• Organizes media libraries
• Maps images to quiz questions
• Maintains hierarchical tagging
• Returns recommendation confidence scores
Semantic image search and recommendations
The platform combines computer vision with vector search to understand both image content and quiz context.
When content creators prepare quizzes, polls, or articles, the system analyzes the content, identifies business intent, and returns the five most relevant images.
Recommendation workflow includes:
• Understand quiz or poll context
• Search newly uploaded images first
• Search existing media library when required
• Rank results using semantic similarity
• Return Top 5 recommendations through APIs
Content teams no longer search for folders manually. Instead, they receive context-aware recommendations directly inside the CMS.
Bulk media automation
The platform automates race weekend media processing from upload through recommendation.
Key experience improvements
• ZIP file processing
• Automatic image categorization
• Metadata enrichment
• Semantic image search
• AI powered recommendations
• Searchable media library
• Independent metadata repository
• API driven CMS integration
This allows Dropt to process large media batches without manual sorting while making every uploaded image immediately searchable.
Infrastructure and deployment
The platform uses a scalable AWS architecture:
• Amazon Bedrock
• Claude 4.5 Sonnet
• Amazon Rekognition
• Amazon S3
• Amazon RDS
• Python
• FastAPI
• AWS API Gateway
• AWS Lambda
• Amazon EC2
• Vector Database
Quality assurance
Validation focuses on recommendation quality, search accuracy, and platform performance.
• Metadata validation
• Recommendation accuracy testing
• Semantic search validation
• Bulk upload testing
• API validation
• Concurrent performance testing
• CMS integration testing
• Real Formula One dataset validation
Impacts
• Up to 90% image categorization accuracy
• 80% significant reduction in manual image sorting
• 2x faster sports campaign creation
• AI powered image management system for scalable media operations
• Semantic search across thousands of sports images
• Intelligent image recommendations integrated into the CMS
About
Before Gen AI and after Gen AI
"With Dropt, AI powered image management system transforms race weekend media into an intelligent content library where every image is automatically organized, searchable, and recommended for the right fan engagement experience."
Prashanna Rao, Head of Engineering, GoML.
Key takeaways for sports media platforms
Common challenges
• Growing media libraries increase manual effort
• Image discovery slows campaign creation
• Content teams spend excessive time organizing assets
Practical guidance
• Deploy AI powered image management system to automate media operations
• Combine computer vision with large language models for richer image understanding
• Use vector search for natural language image retrieval
• Integrate recommendation APIs into existing CMS platforms
• Build scalable media intelligence solutions on AWS
Ready to build AI powered image management system?
Partner with GoML to build AI powered image management system with AI Matic.
