The big new at AWS re:Invent this week was Amazon Q. During his keynote to open the event Tuesday, November 28, Adam Selipsky, CEO, AWS unveiled Amazon Q, described as a new type of generative AI assistant, “that’s an expert in your business.” The value proposition for Amazon Q, Selipsky explained, is to bring generative AI to every builder, cloud engineer, line of business, and specialist in a business. Amazon Q empowers users to ask questions, get answers and – perhaps most importantly – take actions as a result of prompts.
When I had sat through an analyst pre-briefing on the coming announcements on Monday, November 27, I had immediately begun imagining the possible use cases for Amazon Q within AWS’s contact center solution, Amazon Connect. Within minutes, Matt Wood, vice president, technology, AWS, was describing Amazon Q in Amazon Connect.
What Amazon Q Is Designed to Be
Before we discuss how Amazon Q applies in the contact center, let me provide some color into what Amazon Q is designed to be. As has been well documented, the first Gen AI assistant, ChatGPT, was trained on data on the public internet. As a result, Selipsky explained during his Tuesday keynote, ChatGPT or any other similarly trained large language model assistants, don’t know your data, your customers, or your operations. This lack of context limits how useful their suggestions can be.
As seen in a slide presented by Selipsky (above), Amazon Q is trained on 17 years of AWS knowledge. In addition, it connects to 40+ popular data sources, including Amazon S3, Salesforce, Google Drive, Microsoft 365, ServiceNow, Gmail, Slack, and Zendesk. As a result, it can deliver accurate and relevant answers to an enterprise’s business questions, securely and privately. It is secure because it respects existing access controls – Amazon Q only returns information the requestor is authorized to see based on their role.
Many of the examples of Amazon Q described by Selipsky and others throughout the conference were targeted at software developers and other types of IT professionals. Using Amazon Q, customers can ask questions to learn about AWS capabilities (e.g., “Tell me about Agents for Amazon Bedrock?”), research how an AWS service works (e.g., “What are the scaling limits on an Amazon DynamoDB table?”) or figure out the best way to architect a solution (e.g., “What are the best practices for building event-driven architectures?”). Business analysts and executives, using Amazon Q in QuickSight, can access generative AI-powered capabilities to build dashboards and more easily use existing dashboards to simplify decision making.
Amazon Q in Amazon Connect
Let me start by saying the Amazon Connect team has been building and previewing generative AI features for the product throughout 2023. That work did not begin with the announcement of Amazon Q. In a blog post this week, the Amazon Connect team reviews all of the new features in Amazon Connect, including native two-way SMS capability, in-app, web, and video capabilities, and Gen AI-powered post-contact summarization.
As discussed in conversations with Pasquale DeMaio, vice president, AWS, and other Amazon Connect product and marketing leaders, Amazon Q in Amazon Connect can best be thought of as the next generation of an existing Amazon Connect feature, Amazon Connect Wisdom, As I described at the time Wisdom was initially announced - at AWS re:Invent in 2020 - Wisdom is a combination of knowledge base-like capabilities with the ability to pull out relevant inferences from an on-going conversation — from the real-time feature of Amazon Connect Contact Lens — to enable delivery of information to agents that they need, right away.
As seen in the slide below, Amazon Q in Amazon Connect, available now, today is a tool for agents that extends Wisdom’s capabilities. Amazon Q uses real-time customer conversation as the generative AI prompt. It then gives agents suggested responses, next best actions and links to relevant documents and articles. Agents can ask follow-up questions using natural language to better support customers during the interaction.
During a session during the conference, DeMaio reported that his team will continue to iterate Amazon Q in Amazon Connect to improve it. “We already have customers in production with it and they are seeing significant, measurable improvements.”
I’ll close with an insight offered by Werner Vogels, vice president and CTO, Amazon, during his re:Invent keynote on Thursday. “The new AI does not invalidate the old AI… not everything needs to be done with massive, large language models,” he stated. Contact center vendors have been offering features that incorporate artificial intelligence for years. Deployments of these features are not automatically made obsolete by generative AI.
Think of generative AI as creating two possibilities – easier deployment and use of existing AI applications and creation of a new breed of applications that were not previously possible, like Amazon Q.
