How Google's AI-Based Product Discovery Solutions Is Upping The Retail Game

Google Cloud has announced a new suite, known as Product Discovery Solutions for Retail. The new suite is designed to help retailers improve their e-commerce capabilities and deliver highly personalised consumer experiences.

The retail industry is busy migrating to cloud platforms for the past decade. However, as online shopping is gaining momentum, retailers face issues in incorporating improvements in e-commerce tech.

"As the shift to online continues, smarter and more personalised shopping experiences will be even more critical for retailers to rise above their competition," said Carrie Tharp, vice president of retail and consumer at Google Cloud. 

"Retailers are in dire need of agile operating models powered by cloud infrastructure and technologies like artificial intelligence and machine learning to meet today's industry demands. We are proud to partner with retailers around the world, and bring forward our Product Discovery offerings to help them succeed," she added.

Google Cloud's Product Discovery Solutions for Retail is a compilation of the state-of-the-art AI algorithms coupled with the scalable infrastructure of Google Cloud platform. The suite includes the following:

Recommendations AI

In 2019, Google launched Recommendations AI in beta version at the Cloud Next conference. Recommendations AI is a fully managed service, built to assist the retail-oriented businesses to deliver personalised product recommendations to their customers. 

Tech Behind: 

The personalised product recommendation is made possible by sophisticated machine learning algorithms. Recommendations AI uses the latest machine learning architectures of Google with the capability to dynamically adapt to real-time customer behaviour and changes in variables like assortment, pricing, and special offers.

Recommendations AI works in three main steps:

• Ingesting the data
• Customising recommendations
• Delivering recommendations

Pricing: 

Pricing for Recommendations AI is based on operations such as training and tuning (per node per hour) and requesting predictions, such as per 1000 predictions. Also, the prediction cost is tiered by the number of monthly prediction requests.

Know more here. 

Vision API Product Search

The second solution is the Vision API Product Search that allows retailers to create products, each containing reference images that visually describe the product from a set of viewpoints. At present, the solution supports home-goods, apparel, toys, packaged goods, and general.

Tech Behind:

In Vision API Product Search, when the users query the product set with their images, it applies machine learning to compare the product in the user's query image with the images in the retailer's product set. Then, it returns a ranked list of visually and semantically similar results. The solution mainly uses machine learning-powered object recognition algorithms and provides real-time results of similar, or complementary, items from retailers' product catalogue. 

Pricing:

Vision API Product Search pricing is based on the monthly usage for both queries and image management. Also, charges are incurred when the user query a model or maintain an image catalogue via storage.

Know more here.

Google Cloud Search for Retail

Google Cloud Search for Retail offers high-quality product search functionality to retailers and can be embedded into their websites and mobile applications. The solution also provides intuitive search functionalities that are configurable to meet the specific needs of each retail organisation. It is currently available in the Private Preview version 

Wrapping Up

Google Cloud Platform is focused on providing enterprise-grade cloud solutions that leverage Google's cutting-edge technology. With these solutions, retailers can now access the latest technology to gain insights into user intent and context.