Summary Bullets:

• Amazon Web Services (AWS) launched Amazon Q, a generative AI (GenAI) assistant that can be customized using corporate proprietary data, code, and enterprise systems.

• Amazon Q is much more than AWS’ answer to Microsoft’s Copilot, Google’s Duet AI, and OpenAI’s ChatGPT Enterprise; it will be massive for AWS customers.

Amazon has used AWS re:Invent to turbo charge its GenAI strategy with a slew of announcements at every layer of the technology stack, including a new partnership with Nvidia (for more, please see: AWS re:Invent: Amazon and NVIDIA Advance Relationship to Enable Generative AI in the Cloud, November 30, 2023) and new capabilities in its Bedrock portfolio of GenAI services (see AWS Says it Wants to Democratize Generative AI, but Will Amazon Bedrock Live up to Expectations?, April 17, 2023).

However, by far the most salient news is the launch of Amazon Q, an enterprise-oriented GenAI assistant that can be customized using corporate proprietary data, code, and enterprise systems. Amazon Q will help companies enhance internal operations, content creation, decision making, and product development.

Amazon Q is much more than AWS’ answer to Microsoft’s Copilot, Google’s Duet AI, and OpenAI’s ChatGPT Enterprise. It will be massive for AWS customers, providing access to a vast ecosystem of services and applications, while enabling them to apply consistent AI capabilities across the entire platform. Amazon Q will be accessible from the AWS management console. This will enable customers to connect and customize Q with proprietary apps and use over 40 built-in connectors for data repositories including Amazon S3, Dropbox, Confluence, Google Drive, Microsoft 365, Salesforce, ServiceNow, and Zendesk. Amazon claims the chatbot is trained on 17 years of AWS accumulated data and expertise in the cloud space.

For developers, Amazon Q will be transformative in terms of time savings. For example, a capability called ‘code transformation’ will help them migrate applications across technology upgrades at a much faster pace. The chatbot will be integrated into Amazon CodeWhisperer, the company’s machine learning code generator that provides real-time code recommendations. Within a supported IDE (like Amazon’s CodeCatalyst), Amazon Q can generate tests to benchmark software using knowledge of a customer’s code. It can also create a plan and documentation for implementing new features in software or transforming code and upgrading code packages, repositories and frameworks.

Amazon says it has taken strict privacy and safety measures and has been quick to point out that Amazon Q does not use customers’ content to train its underlying models. However, the company has not revealed the underlying technology of Q, which may or may not be based on Anthropic’s Claude 2, the closest technology to GPT-4 available in the market today. Anthropic is also known to follow stricter governance and privacy policies than its famous competitor, OpenAI. Amazon invested around $4 billion in Anthropic in September 2023, providing it with more infrastructure and chips to train and run its models, and the startup’s Claude LLM will power several AWS offerings, including AppFabric.

Currently, Amazon Q is available only for users of Amazon Connect, AWS’ service for contact centers. Eventually, it will be available on other services like Amazon Supply Chain, which helps customers track their supply chain management, and Amazon QuickSight, its platform for business intelligence. Amazon Q for business intelligence is available on preview, and several of its features are free. After the preview phase, it will start at $20 per person per month, and a version with additional features for developers and IT admins will cost $25 per person per month. Amazon Q seems competitively priced compared to Copilot for Microsoft 365 and Duet AI for Google Workspace for the enterprise, with both costing $30 per person per month.