How Spyne is using AI to create visually appealing product catalogs for ecommerce sellers

“To see what a customer buys, you have to see it through a customer’s eyes.” The visual appeal of the products is now influencing a customer’s decision more than ever. If an online seller can showcase their products in the best possible manner, more often than not it results in a conversion.

To cater to matters of product cataloging, Spyne has launched the industry’s first AI-powered, 100 percent fully-automated self-serve technology that lets businesses create studio-finish images and videos at scale within seconds. With Spyne, one can create stunning catalogs 500x faster, at 1/4th cost, driving approximately 40 percent better conversions.

Sanjay Kumar, Co-founder and CEO; Deepti Prasad, Co-founder, Marketing & Branding; Amit Pandey, Head of Engineering at Spyne takes us through the company growth story and how AWS has been a crucial partner in helping build an AI-powered tool.

Growth journey

What started as a visual cataloging company for automobiles, has over the last four years made great progress by venturing into multiple categories. “We saw a great opportunity in automating the end-to-end cataloging process. And with the advancements in AI/ML space, we could solve challenges for product photography space,” says Prasad.

With Ola cars as the first client, Spyne is now serving more than 140+ customers across 15 countries.

Building with AWS

As a deep-tech startup, the core of the company is hugely dependent on AI/ML. The entire beautification of product images is divided into two parts. From guiding the users behind the camera to shoot like a professional photographer to transforming the captured images into something premium, the process is entirely automated.

“We use a lot of computing power from the AWS ecosystem to build these automated models. Amazon SageMaker has helped us train our models and there has been significant cost reduction as well. Similarly, AWS Lambda has led to a reduction in production deployment costs. Overall, we look at AWS as a long-term partner for building cutting-edge solutions at Spyne,” shares Kumar.

Having received $100,000 AWS Activate credits via Accel Partners, Spyne worked closely with AWS account teams to offer AWS services such as Amazon S3, Amazon EC2, Amazon RDS, and AWS Lambda. This allowed the Spyne team to test and evaluate different services without incurring any additional costs.

The impact

“The AWS Solutions Architects team was helpful in terms of migrating our current workload and even some of the internal GPUs. We were able to do all this without any help from internal DevOps teams,” says Pandey.

AWS services have helped the startup scale their training by 10x with multiple GPUs and Amazon SageMaker available at the click of a button thus deploying new models 20-25 percent faster. They have been able to deliver the models faster and have also reduced production costs by 50 percent by spinning up the GPUs as and when required. Moreover, the team has been able to decrease the overall infrastructure costs by 40 percent with the help of the AWS ML pipeline and autoscaling.

The road ahead

Although it started in the automobile category, the plan is to venture into more and more categories including e-commerce, fashion, real-estate etc. “We want to be the de facto cataloging partner for all types of ecommerce businesses and enterprises. By empowering anyone and everyone to shoot with just a smartphone and get stunning consistent catalogs. We aim to drive up the conversions by at least 40 percent at 1/4th the cost and time,” concludes Prasad.