Amazon Q is an artificial intelligence assistant that allows consumers to ask questions about the companies using their data. It was announced during the keynote address given by CEO of AWS Adam Selipsky at AWS re: Invent. To keep an app running well, for instance, an employee can use Amazon Q to learn about the most recent logo usage rules or to decipher another engineer’s code. Q can bring up the relevant information without the worker having to go through piles of paperwork. Soon, other AWS services will be able to use Amazon Q, which is now only available for contact centers.
Amazon says that it’s also building Q its first-party products like QuickSight, a business analytics service. Q within QuickSight can provide visualization options for business reports, automatically reformatting them, or answer questions about the data referenced pic.twitter.com/xtKAnRUs8i
— AI Safety Summit (@AISafetySummit) November 28, 2023
Amazon Q can be accessed through several third-party apps, developer environments (such as Slack), particular company documentation pages, or the AWS Management Console. The questions posed on Amazon Q “will not be used to train any foundation models,” as Selipsky pointed out.
Amazon Q is compatible with all the models available on Amazon Bedrock, an AWS repository of AI models. This includes Claude 2 from Anthropic and Llama 2 from Meta. According to the business, a typical Q user will select a suitable model, link to the model’s Bedrock API, study their data, policies, and process, and finally deploy Amazon Q.
Amazon Web Services (AWS) claims that their AI, Amazon Q, can answer queries about AWS because it was educated using data collected over 17 years. In the case of an Amazon Web Services project, it can recommend the most suitable services. Currently, Amazon Q can only be accessed by contact center customers of Amazon Connect, a service provided by AWS. It will eventually be accessible on other services Amazon provides, such as Amazon QuickSight, a platform for business information, and Amazon Supply Chain, a service that assists clients in monitoring their supply chain management. You may now try out Amazon Q for supply chain management and business intelligence in preview.
![amazon Q](https://b3413276.smushcdn.com/3413276/wp-content/uploads/2023/11/amazon-q-generative-ai-chatbot-656636d7a6795-sej-300x206.webp?lossy=2&strip=1&webp=1)
According to an interview with Dilip Kumar, VP for AWS Applications, which The Verge conducted, each instance of Amazon Q on AWS services would seem different. Q is used in real-time on Amazon Connect to gather information such as account details by listening to customer calls. It saves the contact center representative the trouble of hunting for the right responses.
“We wanted to pair the technology with the services that make the most sense first, and for contact centers, supply chain, and business intelligence, AI is a natural fit,” Kumar said. With Amazon Q in Connect, monthly agent pricing begins at $40. If you visit the AWS Connect website, you can test out Amazon Q “for no charge until March 1, 2024.”
Based on what Selipsky mentioned, Amazon Q can identify customer-set security criteria, preventing unauthorized employees from accessing restricted data through the query system. There are competing products on the market. For Windows users, there’s Microsoft’s Copilot; for document storage, there’s Dropbox’s Dash; and last month, Notion unveiled an AI-powered notes search function.
Selipsky announced that AWS would provide Bedrock customers with the option to create AI-powered apps with guardrails around their models, which coincides with the debut of Amazon Q. Now that the preview is available, the barriers allow businesses to make sure their apps and the models that drive them adhere to their data privacy and responsible AI policies.
Companies, especially those in highly regulated sectors like healthcare and banking, frequently claim they can’t protect their data from being used to train new models. Although it is not currently available, AWS has promised that the guardrails would also include the capability to redact personally identifiable information of its clients’ end users.
Want to know how to get started? See HERE!