Recapping re:Invent: What’s New in Retail

AWS re:Invent 2020, our annual conference for AWS leaders to share the latest advances in AWS technologies, has been a huge success so far. For three weeks, we’ve heard about cutting-edge innovations from keynote speakers, and in this blog, I would like to discuss some of the latest announcements that are most relevant to the retail sector. These new services and features will help retailers streamline store operations, build new customer-aware experiences, optimize demand planning, evolve analytics, enable customer support, and much more.

Simplify Edge Infrastructure

I recently published a blog post on retail edge architecture, which talks about strategies for running compute and machine learning (ML) models at edge locations. re:Invent 2020 has brought some important updates in this space with AWS Panorama and 1U AWS Outposts servers.

AWS Panorama is an ML appliance and software development kit (SDK) that allows retailers to bring computer vision (CV) to existing cameras in their facilities, distributions centers, and other locations to make low-latency predictions locally with high accuracy and low latency. Previously, deploying CV meant building a model, packaging and distributing it, configuring an application to ingest video streams, and then sending data to the cloud to generate insights. The AWS Panorama Appliance streamlines this process and makes it simpler to deploy models, discover camera streams, and start gaining insights. For example, Canadian fuel retailer Parkland Fuel has used the device to gain operational insights into its gas stations, including car counts and wait times at fuel pumps. Retailers can source CV models from AWS and our AWS Partner ISVs or custom build them using Amazon SageMaker, a fully managed service that enables data scientist and developers to build, train, and deploy ML models quickly in the cloud and at the edge.

AWS Outposts 1U and 2U form factors are rack-mountable servers designed specifically for environments such as retail stores, where both physical space and required compute capacity are both smaller than traditional AWS Outposts scenarios, like on-premises data centers. Retailers can now consolidate their existing store applications on a single managed appliance, simplifying IT operations while also establishing a platform for deploying new applications from a centrally managed infrastructure.

Deliver Location-aware Experiences

Amazon Location Service brings geolocation capabilities to application developers, enabling new experiences for shoppers. Take your buy-online-pickup-at-store (BOPUS) strategy to the next level with store geo-fences that automatically dispatch staff to pick and deliver orders as customers arrive. Adding location awareness to your customer engagement strategy can open new omni-channel personalized experiences, such as sending notifications to customers about items in their online shopping carts when they are close to stores that have those specific items in stock. Check out our Amazon Location Service demo at NRF 2021 to learn more.

Optimize Store-level Demand Planning

Amazon Forecast brings the same technology used at Amazon.com to developers as a fully managed service that uses ML to deliver highly accurate forecasts, removing the need for developers to manage application resources or have extensive ML expertise. Many customers use Amazon Forecast for online and in-store demand planning. Now, forecasting just got a little easier with Amazon Forecast Weather Index, a new feature that applies weather data to forecasts to improve prediction accuracy of foot traffic, product sales, and staffing levels by taking into account historical weather patterns at those locations.

Lower Analytics Costs

Many retailers need to consolidate data from various sources to build a complete picture of their customers and operations. This often requires a process that extracts, transforms, and loads (ETL) data into a data lake. We’ve found that the infrastructure to set up and maintain ETL processes can be very expensive and inefficient when not properly optimized. AWS Glue Elastic Views is a new, fully-managed service that allows customers to create materialized views from cloud data stores, such as Amazon DynamoDB, Amazon S3, Amazon Redshift, Amazon OpenSearch Service (September 8, 2021: Amazon Elasticsearch Service has been renamed to Amazon OpenSearch Service), and Amazon Relational Database Service. The service continuously monitors changes in your source data stores and automatically updates the materialized views in your target data stores so that they are always current. It’s serverless and scales capacity up or down automatically based on demand. And since you only pay for what you use, you no longer have to worry about developers leaving expensive ETL clusters running.

Amazon Connect, the AWS contact center as a service, has released several features that can help decrease customer support call times and increase agent productivity and customer satisfaction. When customers call a support line, they typically need to be authenticated, adding time and friction to the experience. But with Amazon Connect Voice ID, customers can opt-in to use their voice for instantaneous identity authentication. Additionally, Amazon Connect Tasks makes it easy to automate, track, and manage agent tasks. It comes with pre-built connectors to support ticketing systems, such as Zendesk and Salesforce, to create a single pane of glass to simplify agents’ work. Amazon Connect Customer Profiles brings together a complete picture of your customers from desperate data sources to lower research time and inefficiencies, and Amazon Connect Wisdom provides ML-powered search and real-time recommendations to help agents resolve issues as quickly as possible.

—

With many more announcements coming out of re:Invent 2020 you can also learn more by visiting What’s New and see a detailed list of retail sessions here. Look for more announcements in the second half of the event, which runs from January 12 – 14, 2021.