Amazon is currently hosting its re:Invent event from November 27 to December 1 in Las Vegas. The company has unveiled numerous new products and services, encompassing new chips, palm-scanning identity services, virtual desktop environments, and more. Nevertheless, the announcement causing a stir on the internet is their introduction of Amazon Q and AI chatbot for AWS customers. Here’s a comprehensive overview of Amazon’s venture into generative AI-powered assistants:
What is Amazon Q?
Amazon Q serves as a natural language query (NLQ) tool, empowering users to pose inquiries about their data using plain English and receive visual responses within seconds. For example, users can input questions like “What is our year-over-year growth rate?” and promptly receive answers through visualisations within QuickSight. This tool is an integral part of the Amazon QuickSight suite, which encompasses various business intelligence (BI) tools and relies on machine learning for its capabilities.
Q utilises a combination of deep learning and machine learning, encompassing natural language processing and schema understanding, to discern users’ intent and comprehend the underlying data, resulting in customised responses. The tool comes pre-trained on diverse data from domains such as sales, marketing, operations, retail, human resources, pharmaceuticals, insurance, energy, and more. This pre-training optimises Q’s ability to comprehend intricate business language, ensuring its effectiveness across various industries and sectors.
Real-world Solutions with Amazon Q
AWS clients set up Q by linking it to and tailoring it with organisation-specific applications and software, including Salesforce, Jira, Zendesk, Gmail, and Amazon S3 storage instances. Q indexes all linked data and content, “learning” various facets of a business, encompassing its organisational structures, core concepts, and product names.
Through a web application, a company can instruct Q to analyse, for example, which product features its customers are grappling with and suggest potential enhancements. Alternatively, akin to ChatGPT, users can upload a file (such as a Word document, PDF, spreadsheet, etc.) and pose queries about that file. Q leverages its connections, integrations, and data, encompassing business-specific details, to formulate responses and provide citations.
Furthermore, Q transcends mere question-answering. The assistant can generate or condense content like blog posts, press releases, and emails. It can also execute actions on behalf of the user through a set of customisable plugins, such as automatically generating service tickets, alerting specific teams in Slack, and updating dashboards in ServiceNow. In order to avert errors, Q prompts users to scrutinise actions before execution and supplies links to the results for validation.
Q is also integrating into Amazon’s contact centre software, Amazon Connect. Now, driven by Q, customer service agents receive recommended responses to customer queries, suggested actions, and links to relevant support articles without manually inputting these customer questions into a text bar. Q additionally produces a post-call summary that supervisors can utilise to monitor subsequent steps.
