Amazon Q (Business and Developer)
Amazon Q is AWS's GenAI assistant family — Q Business (enterprise GenAI over company data, permission-aware) and Q Developer (AI coding assistant). AIF-C01 tests Q Business vs Kendra: Q Business returns synthesised answers; Kendra returns document passages.
Amazon Q was announced at re:Invent 2023 as AWS's answer to Microsoft Copilot and Google Duet. It comes in two distinct products that the AIF-C01 exam treats as separate services.
The Two Products
| Product | What it does | Who uses it |
|---|---|---|
| Amazon Q Business | GenAI assistant that answers questions using company data | Employees, knowledge workers |
| Amazon Q Developer | AI coding assistant for writing, debugging, and explaining code | Developers, DevOps |
Amazon Q Business
A fully managed GenAI assistant that connects to enterprise data sources and answers natural language questions — with citations, permission enforcement, and audit trails.
How it works
- Connect data sources (S3, SharePoint, Confluence, Salesforce, ServiceNow, Jira, RDS, and 40+ connectors)
- Q Business indexes and chunks the content using Bedrock embeddings
- Users ask questions in natural language via the Q Business web UI or embedded widget
- Q Business retrieves relevant chunks, synthesises an answer using a foundation model, and provides source citations
- Answers are permission-aware: users only see content from sources they have access to
Key capabilities
- Synthesised answers with citations — not just document links; generates a coherent response from multiple sources
- Permission-aware retrieval — ACL enforcement means users cannot see content from systems they lack access to (enforced at query time, not just at index time)
- Admin controls — topic blocklist, response guardrails, audit logging to CloudTrail
- Plugins — take actions in third-party systems (create Jira tickets, update Salesforce records) from within Q Business
- Customisation — custom system prompts, response styles, blocked topics
Amazon Q Business vs Amazon Kendra
This is the primary AIF-C01 exam trap:
| Dimension | Amazon Q Business | Amazon Kendra |
|---|---|---|
| Output type | Synthesised GenAI answer with citations | Ranked list of relevant document passages |
| Technology | RAG over FM (Bedrock-powered) | ML-powered semantic search (no FM generation) |
| Permission enforcement | Native — ACL-aware retrieval | Native — ACL-aware retrieval |
| Use case | "What is our PTO policy?" → paragraph answer | "Find documents about PTO policy" → document list |
| Customisation | System prompts, topic blocklist, plugins | Custom document attributes, facets, query tuning |
| Underlying model | Amazon Bedrock FM (selectable) | ML ranking models (not generative) |
Exam trigger for Q Business: "GenAI assistant over company data", "answer questions from SharePoint/Confluence", "synthesise answers from internal docs", "employee self-service AI"
Exam trigger for Kendra: "intelligent document search", "return relevant passages", "enterprise search across multiple repositories", "employees search company policies" (returns passages, does not synthesise)
Amazon Q Developer
An AI coding assistant integrated into IDEs and the AWS Console.
Capabilities
- Code generation — generate code from natural language descriptions
- Code completion — inline suggestions as you type (like GitHub Copilot)
- Code explanation — explain what a block of code does
- Code transformation — modernise Java 8/11 → Java 17/21 (Java upgrade feature)
- Security scanning — detect vulnerabilities in code (SAST-like)
- Unit test generation — generate test cases for selected code
- CLI companion —
/devcommand in AWS CloudShell; Q explains AWS CLI commands - Console integration — asks Q about services in the AWS Management Console
Integrations
- IDE plugins: VS Code, JetBrains IDEs (IntelliJ, PyCharm, etc.), Visual Studio
- AWS Console: Q button in every service page
- AWS CloudShell:
/devnatural language to AWS CLI translation - Amazon CodeCatalyst: integrated throughout the dev workflow
Amazon Q Developer vs GitHub Copilot
| Dimension | Q Developer | GitHub Copilot |
|---|---|---|
| AWS-awareness | Deep — understands AWS services, IAM, CDK | Generic |
| Security scanning | Built-in SAST | Via third-party extensions |
| Java modernisation | Automated Java 8→17 upgrade | Manual |
| Pricing | Free tier (50 inline/month); Pro $19/user/month | $10/user/month |
| IDE support | VS Code, JetBrains, Visual Studio | VS Code, JetBrains, Neovim, etc. |
Exam trigger for Q Developer: "AI coding assistant in the IDE", "generate code for AWS services", "explain code in VS Code", "detect security vulnerabilities in code", "Java modernisation"
Exam Scenario Drill
| Scenario | Service |
|---|---|
| Employees ask HR questions and get paragraph answers from Confluence | Q Business |
| Employees search for documents about a policy and see a list of matching passages | Kendra |
| Developer asks for code completion while writing a Lambda function | Q Developer |
| Company wants a GenAI assistant with SharePoint ACL enforcement | Q Business |
| Developer wants security vulnerability scanning inside VS Code | Q Developer |
| Company wants ML-ranked search results from 10 internal data sources | Kendra |
Key Facts
- Amazon Q Business = GenAI assistant over company data; synthesises answers with citations; permission-aware via ACL enforcement
- Amazon Kendra = enterprise ML search; returns ranked document passages; does not generate answers
- Amazon Q Developer = AI coding assistant; IDE plugins (VS Code, JetBrains); inline completion + security scanning + Java modernisation
- Q Business uses Bedrock FMs under the hood; Kendra uses ML ranking models (not generative)
- Q Business plugins allow actions in third-party systems (Jira, Salesforce) from within the assistant
- Q Developer free tier: 50 inline suggestions/month; Pro tier: $19/user/month
Connections
- cloud/aws-ai-recognition-services — Kendra is in the pre-built AI services family; Q Business is the GenAI successor/complement
- apis/aws-bedrock — Q Business is powered by Bedrock FMs under the hood
- landscape/aws-ai-practitioner — AIF-C01 Domain 3 tests Q Business vs Kendra; Domain 1 covers the RAG concepts Q Business implements
- landscape/aws-cloud-practitioner — CLF-C02 mentions Amazon Q as part of the AI services catalogue
- rag/pipeline — Q Business is a fully managed RAG implementation over enterprise data
- ai-tools/cursor-copilot — Q Developer is AWS's answer to GitHub Copilot; similar positioning
Open Questions
- Is Amazon Q Business converging with Bedrock Agents, or are they on separate roadmaps?
- Does Q Business's ACL enforcement work at query time (re-checks permissions on every query) or at index time only?
Related reading