← Field Guide  ·  Tier 2 · Hyperscaler & Vendor References (Cross-Validation)

AWS

Foundations of Agentic AI on AWS · Bedrock Agents

Hyperscaler · prescriptive guidance4 patterns
In Plain English

This page shows every pattern that AWS has published or written about — 4 in total. Each one includes a simple picture of how it works, a short explanation, an example of when to use it, and its trade-offs (the good parts and the not-so-good parts).

Reading patterns from one source at a time matters because it lets you see how a single company or team thinks about building AI agents, and compare their ideas side by side. If a pattern here also appears in the bigger Deep Catalogue, you will see a link so you can read the full, detailed version.

An open reference book with a pencil in the fold under warm lamp light

2B.01Perceive–Reason–Act Cycle

PERCEIVE–REASON–ACT Agent Perceive Reason Act Observeperceivereasonactobserve

AWS's foundational decomposition of the agent loop: perceive environment → reason about action → act → observe → repeat.

Use caseAll agent designs — a conceptual primitive, not an implementation.
Trade-offsClear mental model · Not prescriptive on mechanism.

Cross-ref: Deep Catalogue PRA-PER-01, PRA-ACT-01, PRA-OBS-01 →

2B.02Bedrock Agent Memory

BEDROCK MEMORY Session Memoryper-conversation Long-Termacross sessions Knowledge BasesRAG-backed

AWS Bedrock-managed short-term and long-term memory stores.

Use caseRecurring user interactions on AWS.
Trade-offsManaged service · Vendor-coupled storage.

See 1C.05

2B.03Bedrock Guardrails

BEDROCK GUARDRAILS Input Input Policy Model Output Policy Safe Output

Managed content filters, PII redaction, topic denials, and grounding checks.

Use caseProduction agents requiring consistent safety policy across models.
Trade-offsManaged and tested · Less customisable than bespoke.

See 1C.14

2B.04Bedrock Agent Orchestration

BEDROCK AGENT Bedrock Agent Action Group A Action Group B Knowledge Base

Bedrock Agents orchestrate action groups (tools) and knowledge bases using a managed planner.

Use caseEnterprise agents on AWS needing low-code setup.
Trade-offsFast to deploy · Platform lock-in.
Also in Tier 2
Microsoft 7Salesforce 4OpenAI 3IBM & Others 2