The blog post explains how to improve the robustness of large language model (LLM) workflows using Amazon Bedrock Guardrails and Amazon Bedrock Agents, demonstrating a use case of an online retail chatbot that should only provide shoe recommendations but can be prompted to give unrelated fiduciary advice.

<div>
<p>This article discusses how to improve the robustness of large language model (LLM) applications using Amazon Bedrock Guardrails and Amazon Bedrock Agents.</p>
<p>Specifically, the article covers:</p>
<ul>
<li>Overview of Amazon Bedrock Agents for building agentic workflows using LLMs</li>
<li>How Amazon Bedrock Guardrails can provide customizable safeguards against harmful or undesired content</li>
<li>A solution architecture using AWS services like Amazon Bedrock, Bedrock Agents, Bedrock Guardrails, Lambda, and SageMaker</li>
<li>Step-by-step notebooks to create and test a retail chatbot agent with and without Bedrock Guardrails</li>
<li>Cost considerations and clean-up instructions</li>
<li>Conclusion highlighting the benefits of using Bedrock Guardrails with Bedrock Agents</li>
</ul>
</div>


Related articles