In a major development, Amazon Web Services (AWS) has launched Amazon Q, a chat tool designed to assist businesses in seeking answers to company-specific queries. AWS CEO Adam Selipsky announced the tool during a keynote speech at AWS re:Invent. Acting as an AI assistant, Amazon Q enables users to pose questions about their businesses, leveraging their data for insights. For instance, employees can inquire about the latest company guidelines for logo usage or seek clarification on a fellow engineer’s code for app maintenance. Rather than manually sifting through numerous documents, Amazon Q surfaces relevant information.
Users can access Amazon Q via the AWS Management Console, individual company documentation pages, developer environments like Slack, and various third-party apps. Selipsky emphasized that the questions asked on Amazon Q ‘will not be used to train any foundation models.’
Amazon Q is compatible with any models available on Amazon Bedrock, AWS’s repository of AI models, including Meta’s Llama 2 and Anthropic’s Claude 2. Customers have the flexibility to choose the model that suits them best, connect to the Bedrock API, learn their data, policies, and workflow, and then deploy Amazon Q. The tool was trained on 17 years’ worth of AWS knowledge, specifically catering to questions related to AWS use. Presently, Amazon Q is accessible only for users of Amazon Connect, AWS’s contact center service, with plans for expansion to other services in the future, such as Amazon Supply Chain and Amazon QuickSight, both currently in preview.
Dilip Kumar, vice president for AWS Applications, explained that each instance of Amazon Q on AWS services would appear different. For Amazon Connect, Q operates in real-time, listening in on customer calls to provide account details and relevant answers to contact center agents without manual searches. The pricing for Amazon Q in Connect starts at $40 per agent per month, with a trial period available until March 1, 2024.
Emphasizing security, Selipsky highlighted that Amazon Q recognizes security parameters set by customers, preventing unauthorized access to sensitive information. Other companies, including Microsoft with Copilot, Dropbox with Dash, and Notion with an AI-powered notes search feature, have introduced similar products. In addition to Amazon Q, AWS announced the availability of guardrails for Bedrock users, offering control over the models used to build AI-powered apps. This preview feature ensures applications and models comply with data privacy and responsible AI standards.