Generative AI and Reinventing with Amazon Q

The Journey to Amazon Q

AWS re:Invent 2023 recently concluded, and there has been some exciting news around data, infrastructure, artificial intelligence, and machine learning. One of the big announcements was Amazon Q, currently under Preview release. Amazon Q is a generative artificial intelligence (AI) powered conversational assistant.

Announced during the keynote by AWS CEO Adam Selipsky, he spoke about various things that AWS is doing to support its customers. The value of constant reinventing – global enterprises, over 80% of unicorns (companies with a Billion-dollar valuation), community colleges, government agencies, non-profits, and various others use AWS heavily in running their business. This heavy usage is the foundation of reinventing with Generative AI. So, how exactly is AWS reinventing with Generative AI?

1. Experience with Artificial Intelligence as a Company

Amazon is not entirely new to experimenting with Artificial Intelligence. Internally, they have been building around AI in optimization, better search capabilities, and delivering new experiences. This knowledge and using it themselves allowed the company to understand the importance of its broad capabilities, best price performance, and consciousness of security.

2. Learning The Stack of Generative AI

Generative AI is the next step in Artificial Intelligence, and its main benefit is to convert to productivity gains for its customers. Adam spoke to the Generative AI stack, starting from the AI infrastructure for training & inference, to tools for building with Large Language Models (LLMs) and other Foundational Models (FMs), to applications leveraging these tools that anyone can build off of even without having the foundational knowledge of the AI infrastructure.

3. Partnerships with Companies

The value of AI comes with speed and data. In order to utilize its full value, the underlying infrastructure needs to be fast. AWS’ partnership with Nvidia enabled the company to make use of the speed that their technology provides. AWS also understands the value of data to make AI more powerful. Throughout the years of research with AI, AWS has partnered with multiple companies that are building AI systems, such as Anthropic and Meta.

4. Amazon CodeWhisperer

AWS understands the value of speed, not only in infrastructure performance but also in increasing the productivity of its builders. Aside from constantly ramping up its training platforms and providing free training to deep cloud knowledge, AWS also worked on tools like CodeWhisperer to provide developers the capabilities to assist in code and focus more on learning the business of the applications they are working on.

5. AI in Chat

In this journey, we now come to AI chat applications and how they have become very valuable in expanding the use of AI. The challenge with the current AI chat applications is they are based on general knowledge and don’t really work on work applications. It has no knowledge of the company, the company’s data, customers, and operations and is a limitation on how it can be used for work productivity; more importantly, there is a limitation of AI knowing its user and how it can limit access to the responses it can provide. This is what Amazon Q aims to solve and is the secret sauce as it enters the field of Generative AI in chat.

What is Amazon Q?

Amazon Q is a new type of generative AI-powered assistant specifically designed for work. It is powered by Amazon Bedrock and implements automated abuse detection, thus taking advantage of the controls implemented in Amazon Bedrock that enforce security and responsible use of artificial intelligence.

Amazon Q is a powerful assistant in using AWS. It can provide service recommendations and an easy search capability for documentation by providing references to its responses. It can also help analyze your current workloads, such as determining misconfigured resources to accelerate building in AWS. It can also help provide guidance on the latest AWS solutions and capabilities. Like any powerful generative AI chat application, it uses the context of the current conversation, allowing the user to ask follow-up questions.

There are various ways to access Amazon Q: through the AWS console, via IDE, CodeCatalyst, Chat (such as Slack or Teams), and soon on mobile. As users log in to their AWS Management Console, a new button will be available on the right side to access Amazon Q. It is intuitive and can provide instructions on how to use it. Several other AWS services (AWS Chatbot, Amazon CodeCatalyst, Amazon Connect, Amazon QuickSight, and Reachability Analyzer) allow for an integration with Amazon Q.

On security, Amazon Q limits usage through IAM permissions. It supports identity-based policies, principal permissions, temporary credentials, and policy actions. No user can gain access to prohibited data within AWS without the right access. It is integrated with AWS CloudTrail and can find request details made to Amazon Q, such as the IP, user, date and time, and other additional details.

To further support users in AWS, Amazon Q can also be utilized to get AWS support. But as earlier stated, it understands a user’s limits, so if you attempt to get AWS support with a Basic Plan, it will instead recommend other options to help you within your limits.

Final Notes

Amazon Q is currently released in Preview mode and is subject to change. With Amazon Q, AWS made it easier for companies to make use of Generative AI with less management overhead while maintaining security around its use. As data becomes more important in driving results, Amazon Q can be utilized for its analysis and application while being more focused on what is available to a company and relevant to its decision-making process.

During the preview period, many of the Amazon Q capabilities are provided without charge. Once it becomes generally available, it will be offered under two pricing plans: Business, and Builder, and has a per-user-per-month pricing model.

References:

AWS re:Invent 2023 – CEO Keynote with Adam Selipsky

https://aws.amazon.com/q/

https://docs.aws.amazon.com/amazonq/latest/aws-builder-use-ug/what-is.html

https://docs.aws.amazon.com/amazonq/latest/business-use-dg/what-is.html