Skip to main content

Command Palette

Search for a command to run...

Amazon Q Business Overview

Updated
5 min readView as Markdown
Amazon Q Business Overview
V

Experienced Software Engineer with 4+ years of expertise, specializing in cutting-edge technologies.

Customers faced issues related to the following, which highlighted the need for an AI powered enterprise assistant:

  1. Information overload affects employee productivity.

  2. There is a vast amount of information to consume.

Amazon Q Business is a fully managed, generative AI-powered assistant with security and privacy for enterprises. It can be utilized for Q&A, summarization, and content creation based on enterprise data.

It is tailored to work with your ERP, Intranet, policies, and documents. It can perform tasks like troubleshooting issues, content generation, working on code, and providing insights. These capabilities can enhance the productivity of employees within the organization. Currently in Preview mode.

Benefits

  1. Accelerate implementation for customers.

  2. Extend and assist human resources in efficiently consuming data.

Challenges for Generative AI Adoption

  1. Accuracy: Even applications built on RAG encounter accuracy issues. Generative AI applications must genuinely comprehend ERP, CRM (e.g., SAP) data to properly process and interpret it.

  2. Security & Control – Enforcing authorization and providing answers based on organizational data, taking into account an individual's access level within the organization. Avoiding false information (Hallucinations) if data is missing with LLM. Ensuring privacy and control over organizational data.

  3. Limited resources – Skilled engineers are essential to handle large volumes of data and effectively utilize generative AI. Enhancing the value of the training engineers, i.e., the investment the organization receives.

Amazon Q areas of expertise

  1. Build trust – Employees trust the generative AI response because it's built on organizational data. For instance, if a user prompt requires security policies from the organization like IAM, the user can rely on the output as it will be generated using the organization's internal data.

    1. Trusted answers generated from enterprise data

    2. In-context conversation

    3. Source references for quick verification

  2. Automate Tasks – Amazon Q not only generates content but also automates tasks. For example, it can create a JIRA ticket. All the plugins listed below are part of the Amazon Q product.

    1. Zendesk

    2. Salesforce

    3. JIRA

Amazon Q Use Cases

  1. Streamline search experience –Traditional approach involves browsing Google to find and process relevant data. With Gen AI, we can get responses and summaries using NLP. When data is gathered from various sources, it will also offer citations.

  2. Accelerate content creation – When creating content, we use a Gen AI-generated template 60% of the time.

  3. Generate summary – It can create a summary of large data, for example, team meeting call summaries or summaries of anything relevant to me.

  4. Extract key insights – Utilize visuals such as pie charts to extract important information.

Amazon Q Key Features

  1. Upload files and analyze content.

    1. The admin can determine which users are allowed to upload files.

    2. Basic safety guardrails are already in place by default.

    3. Nightly or periodic syncing with the data source is supported.

  2. Execute actions based on plugins. The user experience for plugins includes creating a JIRA ticket within chat or accessing the past history of JIRA tickets per user by pulling data. The following plugins are supported:

    1. Zendesk

    2. Salesforce

    3. JIRA

    4. ServiceNow

  3. Safety and security are top priorities. The system is designed to acknowledge enterprise user permissions. It can integrate with SAML 2.0 compliant IDPs and ADs (such as Keycloak – an Open Source Identity and Access Management solution, and AWS service SAML 2.0).

  4. Adhere to data privacy and security checks i.e.

    1. Follow information security and norms.

    2. Follow ethical norms.

    3. Avoid creating conflicts.

By default, there are basic guardrails in Amazon Q:

  1. Violence and terrorism-related queries will be filtered.

  2. Malicious content will be identified.

  3. Organization-level blocked words.

  4. Confidential information will be protected, meaning unauthorized users are not allowed to access confidential organization-level information.

  5. Guardrail to allow users to upload PDFs.

  6. Responses are restricted based on RAG. Only provide answers using RAG knowledge. For example, for the query "What is the capital of France?" If the relevant data is not in RAG, it won't provide an answer even if it's in LLM knowledge.

  1. Explicit guardrails include:

    1. Respond with predefined message.

    2. Restrict responses to enterprise data.

    3. Restrict responses to enterprise content.

    4. Apply guardrails to specific users and groups in the enterprise.

Amazon Q Architecture

Image credits - Amazon Web Services

DB, Salesforce, SharePoint, Drive, S3 bucket, etc., are used by Amazon Q business to ingest document content. For permissions information, it interacts with the Identity Provider for user and group info. Users interact with Amazon Q to make queries and receive permission-filtered responses.

Business Benefits

  1. Save costs – IT helpdesk (e.g., calls for password resets).

  2. Save time – quicker employee onboarding for certain projects.

  3. Improved customer service - Amazon Q can handle customer inquiries and resolve issues.

  4. Enhance team collaboration - seamless collaboration and communication among team members

  5. Boost productivity - Automate tasks and workflows.

Case Studies

Following section mentions some of the case studies related to Amazon Q business.

  1. Gilead Sciences, Inc. – a pharmaceutical company involved in clinical research, data pipelines, and data engineering. By using Amazon Q, they extracted insights and sped up the analysis of large amounts of Gilead data across the enterprise.

  2. Wunderkind - a top digital marketing platform, reduced time spent on content discovery by over 30% and accelerated the content creation process by almost 50% using Amazon Q business.

Pricing Plans

Amazon Q Business offers different tiers of user subscriptions and index types. Pricing is around $20/ month per user (SAML user) and also depends on documents ingested.

User subscriptions

  • Amazon Q Business Pro subscriptions - suitable for knowledge workers, enhancing productivity across various tasks.

  • Amazon Q Business Lite subscriptions - designed for company-wide deployment to all employees. It is used for scenarios like IT help desks and other Q&A chatbot use cases.

Index capacity

  • Enterprise index - The Enterprise index is ideal for production workloads. It is spread across 3 Availability Zones for fault tolerance.

  • Starter index - The Starter index is for proof-of-concept or developer workloads. It is set up in a single Availability Zone.

Image credits -https://aws.amazon.com/q/business/pricing/

Image credits -https://aws.amazon.com/q/business/pricing/

References

More from this blog

Untitled Publication

8 posts