<?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/fde-si-inhouse-delivery-model-comparison</loc>
    <news:news>
      <news:publication>
        <news:name>삼태연구소</news:name>
        <news:language>ko</news:language>
      </news:publication>
      <news:publication_date>2026-09-22T12:30:36.334Z</news:publication_date>
      <news:title>FDE·전통 SI·인하우스팀 비교: AI 프로젝트 수행 모델을 책임 구조로 고르는 법</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://samtaelabs.com/blog/multi-location-booking-management-standardization-guide</loc>
    <news:news>
      <news:publication>
        <news:name>삼태연구소</news:name>
        <news:language>ko</news:language>
      </news:publication>
      <news:publication_date>2026-09-22T03:00:28.145Z</news:publication_date>
      <news:title>여러 지점 예약 관리 시스템 통합, 본사와 지점이 먼저 같게 정할 것</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://samtaelabs.com/blog/ai-human-approval-boundaries-requirements-design</loc>
    <news:news>
      <news:publication>
        <news:name>삼태연구소</news:name>
        <news:language>ko</news:language>
      </news:publication>
      <news:publication_date>2026-09-21T22:30:03.556Z</news:publication_date>
      <news:title>AI 기능이 실패하는 지점은 모델이 아니라 승인 경로다: 사람의 책임을 남기는 요구사항 설계법</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://samtaelabs.com/blog/existing-app-maintenance-transfer-readiness</loc>
    <news:news>
      <news:publication>
        <news:name>삼태연구소</news:name>
        <news:language>ko</news:language>
      </news:publication>
      <news:publication_date>2026-09-21T04:51:34.082Z</news:publication_date>
      <news:title>기존 앱을 다른 개발사에 유지보수 맡기기 전, 고칠 수 있는지 확인하는 법</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://samtaelabs.com/blog/open-weight-ai-model-enterprise-evaluation-criteria</loc>
    <news:news>
      <news:publication>
        <news:name>삼태연구소</news:name>
        <news:language>ko</news:language>
      </news:publication>
      <news:publication_date>2026-09-21T03:00:28.172Z</news:publication_date>
      <news:title>오픈 웨이트 AI 모델, 기업 도입 전에는 무엇을 비교해야 하나: 상업 사용·수정·재배포·감사 기준</news:title>
    </news:news>
  </url>
  <url>
    <loc>https://samtaelabs.com/blog/computer-use-agent-platform-evaluation-isolation-driver-reproducibility</loc>
    <news:news>
      <news:publication>
        <news:name>삼태연구소</news:name>
        <news:language>ko</news:language>
      </news:publication>
      <news:publication_date>2026-09-20T22:30:03.575Z</news:publication_date>
      <news:title>컴퓨터 사용 에이전트 도입 판단법: 모델 성능보다 격리 환경·드라이버·재현성을 비교하라</news:title>
    </news:news>
  </url>
</urlset>