Amazon Web Services (AWS) CEO Adam Selipsky announced a new product called Amazon Q at the AWS re:Invent conference on Tuesday, which tries to improve upon Microsoft’s Copilot assistant.

Amazon Q is a generative AI-powered assistant that can help users with various tasks at work, such as chatting, generating content, and taking actions. And from Selipsky’s description, it does seem as comprehensive as Microsoft’s assistant. (The announcement was part of Selipsky’s keynote, where he made several other announcements. Highlights are here.)

The Q assistant lets employees ask questions using natural language, and can understand and access the user’s company systems, data repositories, and operations. Selipsky emphasized the importance of security and privacy, and said that Amazon Q respects the user’s existing identities, roles, and permissions. He also said that Amazon Q does not use the user’s content to train its underlying models.

Selipsky said Amazon Q will be transformative for different kinds of people who do different kinds of work, and gave some examples of how Amazon Q can help developers and IT professionals with building and managing applications on AWS.

An expert assistant for building on AWS

According to Selipsky, Amazon Q has been trained with 17 years worth of AWS knowledge, and can provide guidance and recommendations on AWS capabilities, technologies, solutions, best practices, and implementations. Users can chat with Amazon Q in the AWS Management Console, documentation, IDE, or team chat rooms like Slack.

For example, users can ask Amazon Q questions about how to build a web application with AWS, what their options are, and how to get started. Users can also follow up with more questions to narrow down their options or get more details. Selipsky said that Amazon Q can also help users with troubleshooting and optimizing their workloads. Users can also ask Amazon Q to analyze their network configuration and connectivity issue, saving customers a lot of time and effort with these tasks, Selipsky said.

A creative assistant for generating content

Selipsky also said Amazon Q can help users with generating content for various purposes, such as writing code. Users can simply tell Amazon Q what they want to create, and Amazon Q will generate the content for them.

Selipsky said that Amazon Q can also help users with adding new features to their applications. Users can just write a prompt, and Amazon Q will create a draft plan, which users can then collaborate and improve using natural language chat. Then, Amazon Q will implement the plan across the files in the application, and users can review the changes and ensure the quality. Selipsky said that this feature is available today in the IDE.

A smart assistant for upgrading and maintaining code

Selipsky also demonstrated how Amazon Q can help developers with upgrading and maintaining their code, to avoid tedious and time-consuming tasks. He said that Amazon Q has a code transformation capability, which can help users with language version upgrades, framework migrations, security patches, and more.

Users can simply ask Amazon Q to perform the code transformation, and Amazon Q will handle the rest. It will identify the necessary changes, replace the deprecated code, incorporate the best practices, and run the tests on the upgraded application. Selipsky said that Amazon Q can do this in a fraction of the time that it takes today.

Selipsky shared an example of how Amazon Q was used internally to upgrade 1000 applications from Java 8 to Java 17 in just two days. He said that Amazon Q currently supports Java upgrades, but will soon support transformations for other languages and frameworks as well, as .Net migrations to Linux.

A business expert for finding and analyzing information, including for BI and customer support

Selipsky also said Amazon Q can be a business expert, by connecting to over 40 popular enterprise systems, such as S3, Salesforce, Microsoft, Google, Slack, and more. He said that Amazon Q can index and learn from the user’s data and content, and use generative AI and semantic search to help a user find relevant content. 

The presentation showed how Amazon Q displays responses to business questions using tables, charts, and bullet points. Amazon Q provides links to the sources and documents it uses, and allows the user to drill down into more details or ask follow-up questions. He said that Amazon Q can do all this in just a fraction of a second, using the power of AWS.

Finally, Silepsky showed how Q can connect to business intelligence applications using QuickSite, and connect to Amazon’s customer service center application, Amazon Connect. He said Q can use a customer call itself as a prompt, which triggers answers from Q that a call agent can use without even typing in requests.