<?xml version="1.0" encoding="UTF-8"?>
<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9"
        xmlns:news="http://www.google.com/schemas/sitemap-news/0.9">
  <url>
    <loc>https://samtaelabs.com/blog/ai-agent-fault-tolerant-architecture-requirements-design</loc>
    <news:news>
      <news:publication>
        <news:name>삼태연구소</news:name>
        <news:language>ko</news:language>
      </news:publication>
      <news:publication_date>2026-10-09T22:30:03.595Z</news:publication_date>
      <news:title>AI 에이전트 도입 시 장애 복구 아키텍처를 요구사항 단계에서 설계해야 하는 이유</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://samtaelabs.com/blog/ai-agent-vendor-selection-real-web-compatibility-failure-patterns</loc>
    <news:news>
      <news:publication>
        <news:name>삼태연구소</news:name>
        <news:language>ko</news:language>
      </news:publication>
      <news:publication_date>2026-10-09T03:00:28.269Z</news:publication_date>
      <news:title>AI 에이전트 벤더를 고를 때, 마케팅 데모보다 먼저 봐야 할 것들</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://samtaelabs.com/blog/llm-rag-data-source-trust-requirements-design</loc>
    <news:news>
      <news:publication>
        <news:name>삼태연구소</news:name>
        <news:language>ko</news:language>
      </news:publication>
      <news:publication_date>2026-10-08T22:30:03.145Z</news:publication_date>
      <news:title>LLM·RAG 도입 시 데이터 출처 신뢰성을 요구사항 단계에서 설계해야 하는 이유</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://samtaelabs.com/blog/domain-asset-first-ai-product-planning</loc>
    <news:news>
      <news:publication>
        <news:name>삼태연구소</news:name>
        <news:language>ko</news:language>
      </news:publication>
      <news:publication_date>2026-10-08T03:00:28.002Z</news:publication_date>
      <news:title>AI 기능을 붙이기 전에, 도메인 자산부터 목록으로 만들어야 하는 이유</news:title>
    </news:news>
  </url>
</urlset>