--- title: I Built My First AI Agent With AWS AgentCore, and the... description: A learner recounts building a customer support AI agent using Amazon Bedrock AgentCore and the Strands SDK as part of Udacity's AWS Agent Engineer Nanodegree.... canonical: https://daily.dev/posts/i-built-my-first-ai-agent-with-aws-agentcore-and-the-hardest-part-wasn-t-the-ai-z5jrclpu6 twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: I Built My First AI Agent With AWS AgentCore, and the Hardest Part Wasn't the AI | daily.dev og:description: A learner recounts building a customer support AI agent using Amazon Bedrock AgentCore and the Strands SDK as part of Udacity's AWS Agent Engineer Nanodegree.... og:url: https://daily.dev/posts/i-built-my-first-ai-agent-with-aws-agentcore-and-the-hardest-part-wasn-t-the-ai-z5jrclpu6 og:image: https://api.daily.dev/og/posts/z5jrcLPu6.png og:image:alt: I Built My First AI Agent With AWS AgentCore, and the Hardest Part Wasn't the AI og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # I Built My First AI Agent With AWS AgentCore, and the Hardest Part Wasn't the AI **[DEV](https://daily.dev/sources/devto)** · 21 min read · 1 upvotes · 0 comments ## Summary A learner recounts building a customer support AI agent using Amazon Bedrock AgentCore and the Strands SDK as part of Udacity's AWS Agent Engineer Nanodegree. The agent handled order tracking, refunds, knowledge-base Q&A, cross-session memory, loyalty discount calculations via Code Interpreter, and live web browsing via a Browser Tool, wiring together AgentCore Runtime, AgentCore Gateway, Lambda, API Gateway, IAM, and CloudWatch. Key lessons include treating deployment as the start of testing rather than the end, debugging permissions before assuming code bugs, separating model reasoning from precise code-based calculations, and cleaning up cloud resources after finishing. The main takeaway is a mindset shift from trying to understand an entire architecture at once to tackling one piece at a time. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### What AWS services do you need to connect to build a customer support AI agent with AgentCore? A working setup connects AgentCore Runtime for hosting the deployed agent, AgentCore Gateway to link the agent to backend tools, API Gateway and Lambda for order and refund operations, a Bedrock Knowledge Base for product and policy retrieval, AgentCore Memory for cross-session customer context, Code Interpreter for precise calculations, and a Browser Tool for live web interaction, all monitored through CloudWatch. _Developers piecing together multi-service agent architectures can find real build breakdowns like this on daily.dev._ ### Why did my AgentCore Browser Tool fail to start a browser session even though it was configured? A misconfigured or missing IAM permission on the AgentCore runtime is a common cause, not a broken Browser Tool. In one build, the runtime lacked the permission needed to start a browser session; adding the correct permission resolved it immediately, without any code changes. _Anyone debugging AgentCore permission issues can compare notes on daily.dev before assuming the tool itself is broken._ ### Why did my AI agent's refund confirmation not match what actually happened in the backend? A model stating that a refund was processed does not guarantee the backend action actually succeeded; the agent's confirmation text and the real system state are two separate things. The fix is to trigger the refund through an actual backend operation (via AgentCore Gateway and Lambda) and have the agent respond based on the real result, not an assumed success. _Teams building agents that trigger real actions can share verification patterns like this on daily.dev._ ## Similar posts on daily.dev - [Build and Run Your Own AI Agent in the Cloud](https://daily.dev/posts/build-and-run-your-own-ai-agent-in-the-cloud-aoostzrec) · Towards Data Science · 1 upvotes · 0 comments - [From 'Works on My Machine' to Production-Ready: Building AI Agents with Amazon Bedrock AgentCore](https://daily.dev/posts/from-works-on-my-machine-to-production-ready-building-ai-agents-with-amazon-bedrock-agentcore-bt6s0repy) · Pulumi · 1 upvotes · 0 comments - [How to Deploy an AI Agent with Amazon Bedrock AgentCore](https://daily.dev/posts/how-to-deploy-an-ai-agent-with-amazon-bedrock-agentcore-ooj3xmxsd) · freeCodeCamp · 2 upvotes · 0 comments - [Build AI-powered dashboard automation agents with NLP on Amazon Bedrock AgentCore](https://daily.dev/posts/build-ai-powered-dashboard-automation-agents-with-nlp-on-amazon-bedrock-agentcore-jgk7lybxg) · AWS · 1 upvotes · 0 comments --- Tags: [#aws](https://daily.dev/tags/aws), [#ai-agents](https://daily.dev/tags/ai-agents), [#rag](https://daily.dev/tags/rag), [#aws-lambda](https://daily.dev/tags/aws-lambda) [View this post on daily.dev](https://daily.dev/posts/i-built-my-first-ai-agent-with-aws-agentcore-and-the-hardest-part-wasn-t-the-ai-z5jrclpu6) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"I Built My First AI Agent With AWS AgentCore, and the Hardest Part Wasn't the AI","url":"https://daily.dev/posts/i-built-my-first-ai-agent-with-aws-agentcore-and-the-hardest-part-wasn-t-the-ai-z5jrclpu6","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/i-built-my-first-ai-agent-with-aws-agentcore-and-the-hardest-part-wasn-t-the-ai-z5jrclpu6"},"datePublished":"2026-09-29T17:46:04.977Z","dateModified":"2026-09-29T19:24:38.198Z","description":"A learner recounts building a customer support AI agent using Amazon Bedrock AgentCore and the Strands SDK as part of Udacity's AWS Agent Engineer Nanodegree....","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/8b6c798ef289ae60fdfb2c056bf3444c?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/8b6c798ef289ae60fdfb2c056bf3444c?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"DEV","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"DEV","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/devto","url":"https://daily.dev/sources/devto"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/i-built-my-first-ai-agent-with-aws-agentcore-and-the-hardest-part-wasn-t-the-ai-z5jrclpu6","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":1},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"aws,ai-agents,rag,aws-lambda","timeRequired":"PT21M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"DEV","item":"https://daily.dev/sources/devto"},{"@type":"ListItem","position":3,"name":"I Built My First AI Agent With AWS AgentCore, and the Hardest Part Wasn't the AI"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/i-built-my-first-ai-agent-with-aws-agentcore-and-the-hardest-part-wasn-t-the-ai-z5jrclpu6#faq","mainEntity":[{"@type":"Question","name":"What AWS services do you need to connect to build a customer support AI agent with AgentCore?","acceptedAnswer":{"@type":"Answer","text":"A working setup connects AgentCore Runtime for hosting the deployed agent, AgentCore Gateway to link the agent to backend tools, API Gateway and Lambda for order and refund operations, a Bedrock Knowledge Base for product and policy retrieval, AgentCore Memory for cross-session customer context, Code Interpreter for precise calculations, and a Browser Tool for live web interaction, all monitored through CloudWatch. Developers piecing together multi-service agent architectures can find real build breakdowns like this on daily.dev."}},{"@type":"Question","name":"Why did my AgentCore Browser Tool fail to start a browser session even though it was configured?","acceptedAnswer":{"@type":"Answer","text":"A misconfigured or missing IAM permission on the AgentCore runtime is a common cause, not a broken Browser Tool. In one build, the runtime lacked the permission needed to start a browser session; adding the correct permission resolved it immediately, without any code changes. Anyone debugging AgentCore permission issues can compare notes on daily.dev before assuming the tool itself is broken."}},{"@type":"Question","name":"Why did my AI agent's refund confirmation not match what actually happened in the backend?","acceptedAnswer":{"@type":"Answer","text":"A model stating that a refund was processed does not guarantee the backend action actually succeeded; the agent's confirmation text and the real system state are two separate things. The fix is to trigger the refund through an actual backend operation (via AgentCore Gateway and Lambda) and have the agent respond based on the real result, not an assumed success. Teams building agents that trigger real actions can share verification patterns like this on daily.dev."}}]} ```