Amazon Web Services Inc. said today it’s trying to embed more intelligence into the industrial sector with a number of new machine learning tools and services aimed at helping customers monitor their workers, the machines and equipment they operate, and the environments they work in.

The new machine learning services, plus a range of others, were introduced during the virtual AWS re:Invent 2020 event in a keynote by Chief Executive Andy Jassy (pictured).

Possibly the most important of the new services is Amazon Monitron. As the name suggests, it’s designed to monitor equipment and machines and send alerts to the engineering team if it shows any signs of breaking down. When industrial firms know their equipment is failing, they can perform maintenance before it does, at a time that suits them best, rather than having to shut down operations at an inappropriate time when it stops working.

Jassy said during his presentation that experienced engineers often know when a piece of machinery is breaking down by a change in its sound or vibration. But teams could benefit from an earlier warning, he said.

“A lot of companies either don’t have sensors, they’re not modern powerful sensors, or they are not consistent and they don’t know how to take that data from the sensors and send it to the cloud, and they don’t know how to build machine learning models, and our manufacturing companies we work with are asking [us] just solve this [and] build an end-to-end solution,” Jassy said.

The Amazon Monitron service includes small, low-cost sensors (pictured here) that can be attached to machines to monitor them for abnormal vibrations and temperatures that might indicate a problem. These are then combined with a machine learning model that understands what a normal state looks like for each machine. It then uses that information to recognize any anomalies that might show up, reporting back to the engineering team through a mobile app to show which machines need maintenance based on the data it sees.

For companies that have newer machines that comes with their own sensors, and therefore don’t need the full Monitron service, they could instead use Amazon Lookout for Equipment. This is a new service that gives customers a way to send their sensor data to AWS and use pre-built machine learning models to detect any abnormal machine behavior, Jassy said.

Amazon Panorama, meanwhile, is a service that uses computer vision technology to analyze video camera footage within industrial facilities in order to detect any issues with safety or compliance. For example, it can spot workers who aren’t wearing the prescribed personal protective equipment in a certain part of the building, or vehicles entering an unauthorized area.

There’s also a separate Software Development Kit being made available to hardware providers to help them build new cameras to be used with Amazon Panorama.

Meanwhile, Amazon Lookout for Vision is a related service that relies on computer vision and machine learning to detect anomalies, with a focus on machines and their parts. So, for example, it might be able to spot a dent in a panel or an irregular shape. It works by comparing the images of machines it sees with initial baseline images that show how things are meant to be.

Analyst Holger Mueller of Constellation Research Inc. said the new industrial services show how Amazon is moving away from its traditional focus on infrastructure and platform services and into software-as-a-service layers.

“Monitron is the perfect example of this,” he said. “It shows the maturation of the basic underlying components in the cloud, which allows AWS to bring together multiple services to help enterprises, in this case by monitoring, logging and maintaining their equipment. This is the year the cloud goes vertical, and Monitron has appeal for manufacturing and machine-intensive industries,” he said.

Smarter contact centers

Also in the realm of machine learning, Amazon announced today a number of new capabilities for its Amazon Connect offering for customer services teams.

Amazon Connect is a service that can spin up a virtual contact center in a few minutes and then automate many aspects of customer service using chatbots and other, artificial intelligence-based tools.

The new capabilities include something called Amazon Connect Wisdom that has built-in connectors to knowledge repositories. It works by listening into calls with customers and then delivering specific information to customer service agents so they can provide a more personalized service, Jassy said.

Also new is Amazon Connect Customer Profiles, which aggregates a customer’s order history, contact history, sales, CRM and other data. Real-Time Contact Lens for Amazon Connect is used to identify customer issues in real-time during a call so agents can respond faster. Amazon Connect Tasks helps to track and manage tasks for agents and supervisors, and Amazon Connect Voice ID provides real-time caller authentication using machine learning.

Tailored recommendations for application developers

Developers look set to benefit from Amazon’s growing expertise in machine learning too. Jassy also found time to announce a new service called DevOps Guru. It’s aimed at DevOps teams made up of developers and information technology staff, and the purpose is to help them find issues with their code that could be having an impact on their application’s performance.

Amazon DevOps Guru can be considered as a sibling to the Amazon CodeGuru service announced last year that helps developers find issues in their code before it gets deployed. DevOps Guru works similarly, using machine learning to find any problems on the operations side, or once the app is up and running.

The service collects and analyzes data from an application’s metrics, logs and events to identify any behavior that deviates from the norm, then provides an alert through SMS, Slack or another service, so DevOps teams can investigate.

This is actually a big deal, because Amazon DevOps Guru means Amazon is now competing in an emerging application observability space that’s currently dominated by firms such as Splunk Inc., Sumo Logic Inc. and DataDog Inc.

Mueller told SiliconANGLE that it’s a new era in which it’s now possible to automate key software development processes, also known as self-driving or autonomous software.

“After launching its CodeGuru offering last year, AWS is now tackling DevOps challenges with DevOps Guru,” Mueller said. “It’s providing more and better automation for enterprises to deliver their code assets to production and to monitor their environments. DevOps professionals might initially be skeptical of this, but applying machine learning to DevOps challenges is the True North for code and environments, as we see with the AIOps trend. It brings us one step closer to the Holy Grail of DevOps, which is autonomous software operations, where software can write, fix, deliver and operate itself.”

Smarter business intelligence

One final machine learning-related update today pertains to Amazon QuickSight, a business intelligence tool that rivals products sold by Tableau Software Inc. and Qlik Inc. Amazon QuickSight provides a way for employees to derive business insights from company data, regardless of their technology expertise. It provides range of graph-based tools that people can use to build data visualizations, charts and tables, using information pulled from AWS or third-party databases.

Amazon QuickSight Q is a new natural language query tool that functions as a kind of companion to AWS QuickSight. It enables users to search through databases using their natural voice and get a response immediately. So, for example, it can recognize terms such as “revenue,” “growth” and “allocation” whenever an employee asks a question, and provide answers in the shape of charts, graphs or other visualizations.

“We will provide natural language to provide what we think the key learning is,” Jassy said. “I don’t like that our users have to know which databases to access or where data is stored. I want them to be able to type into a search bar and get the answer to a natural language question.”

AWS said QuickSight Q is available in preview in the US East (N. Virginia), US West (Oregon), US East (Ohio) and Europe (Ireland) regions.

Image: Quinny Pig/Twitter