Amazon’s cloud computing division Amazon Q is a chat service that AWS introduced that allows organizations to ask inquiries about their own operations.
AWS CEO Adam Selipsky introduced Amazon Q at AWS re:Invent. It functions as an AI assistant that allows users to ask questions about their businesses using their data. Employees can, for instance, ask Amazon Q about the most recent policies on the use of the corporate logo or decipher another engineer’s code to keep an app up to date. Rather than the employee having to go through dozens of documents, Q can retrieve the information.
Through the AWS Management Console, developer environments such as Slack, documentation sites from specific firms, and other third-party tools, users can access Amazon Q.
Questions submitted to Amazon Q “will not be used to train any foundation models,” according to Selipsky.
Any model on Amazon Bedrock, AWS’s collection of AI models, including Claude 2 from Anthropic and Llama 2 from Meta, can be used with Amazon Q. Customers who use Q, according to the firm, often select the model that best suits their needs, connect to the model’s Bedrock API, use it to learn about their data, process, and regulations, and then implement Amazon Q.
Amazon Q, according to AWS, may be used to ask questions about using AWS because it was trained using 17 years of AWS knowledge. For a project, it can recommend the finest AWS services.
At the moment, Amazon Q is exclusive to customers of AWS’s contact center service, Amazon Connect. It will eventually be accessible on more services including Amazon QuickSight, the company’s business intelligence platform, and Amazon Supply Chain, which assists users in tracking their supply chain management. You may preview Amazon Q for business intelligence and supply chain management.
In an interview with The Verge, Dilip Kumar, vice president for AWS Applications, stated that every Amazon Q instance on AWS services will have a unique appearance. Q is implemented in real time on Amazon Connect, essentially listening in on a client call to obtain information such as account details. It provides the contact center representative with pertinent answers to queries so they don’t have to look them up on their own.
“Our goal was to match the technology with the most practical services first, and artificial intelligence (AI) is a perfect fit for contact centers, supply chains, and business intelligence,” said Kumar.
The monthly cost of Amazon Q in Connect begins at $40 for per agent. On the AWS Connect website, customers can test out Amazon Q “for free until March 1, 2024.”
According to Selipsky, Amazon Q understands security settings made by users, meaning that staff members who aren’t authorized to read certain data can’t utilize the query system to obtain information.
Similar items have been created by other firms. Similar functionality is provided by Dropbox’s Dash, Microsoft’s Copilot for Windows users, and Notion’s AI-powered notes search capability, which was unveiled last month.
Selipsky noted that in addition to the introduction of Amazon Q, AWS would allow Bedrock users to impose restrictions on the models they utilize to create AI-powered applications. The guardrails, which are now in preview, enable businesses to guarantee that their models and applications adhere to responsible AI and data privacy requirements.
Businesses frequently refer to their incapacity to fence off their data and ensure that their information is not utilized to train future versions of the model, especially those in highly regulated industries like finance and healthcare. The guardrails will include the ability to redact the personally identifiable information of their clients’ end users, according to AWS, albeit it is not currently accessible.
