TABLE OF CONTENT

About the Client

eJohri is India’s first omnichannel jewelry marketplace, connecting buyers with jewelers across metropolitan, urban, and suburban regions. The platform enables seamless online purchases and facilitates offline discovery. Offering a diverse jewelry range—gold, silver, platinum, solitaire, and more—eJohri enhances accessibility and convenience for both sellers and buyers.

Business Situation

eJohri faced several hurdles in scaling its jewelry marketplace due to inefficient image processing:

Time-consuming manual editing – Each jewelry image required 2-3 hours of manual editing.
Delayed seller onboarding – Slower image processing led to onboarding delays for jewelers.
Inconsistent image quality – Manual editing resulted in varying image quality.
Scalability concerns – As the platform grew, manual processing became unsustainable.
Limited dataset – Lack of a large dataset hindered AI model training.
Need for high accuracy – AI-based editing had to match manual precision.

The Solution

Unthinkable leveraged AI and machine learning to automate jewelry image editing:

To address eJohri’s challenges, Unthinkable initially explored computer vision techniques such as edge detection and unsupervised learning. However, due to accuracy limitations across different jewelry types, the team pivoted to machine learning-based solutions. Using transfer learning, Unthinkable trained a high-accuracy AI model on a generalized jewelry dataset, overcoming the limited data availability issue. This approach enabled the automation of image processing while maintaining quality and precision.

Developed an AI-powered model to automate jewelry image background removal.
Implemented transfer learning to enhance model training with a diverse dataset.
Ensured 95% accuracy to maintain high image quality standards.
Reduced editing time from hours to just a few seconds.
Enabled seamless seller onboarding by accelerating image processing.
Eliminated manual dependency for large-scale jewelry image editing.

The Impact

By implementing Unthinkable’s AI-powered image processing solution, eJohri significantly enhanced its operational efficiency. The AI model automated jewelry image editing, reducing processing time from 3-4 hours to just 2-3 seconds, ensuring rapid seller onboarding and consistent image quality. This innovation saved eJohri hundreds of man-hours and allowed them to scale their platform seamlessly.

Conclusion

Unthinkable’s AI-driven solution transformed eJohri’s image processing workflow, enabling jewelers to onboard faster and sell more efficiently. By automating a previously manual process, the marketplace improved scalability, accuracy, and overall user experience. This case highlights how AI can revolutionize digital commerce by eliminating bottlenecks and enhancing productivity.