{
  "id": 5756853,
  "title": "A Merchant-Controlled Architecture for AI Support",
  "url": "https://urgent.news/2026/09/05/a-merchant-controlled-architecture-for-ai-support",
  "topic": "ai",
  "section": "AI",
  "published": "2026-09-05T10:00:01.000Z",
  "source": {
    "name": "Dev.to",
    "slug": "dev-to",
    "url": "https://dev.to/wukongchat/a-merchant-controlled-architecture-for-ai-support-3kj5"
  },
  "original_language": "en",
  "account": "An effective approach to AI-powered support involves establishing clear operating rules and maintaining transparency for merchants. The key is to define authoritative sources, scope, conditions, and handoff procedures. This contract should distinguish informative replies from operational resolutions, ensuring that AI assistance remains distinguishable from completed support work. By modeling knowledge as maintained data with clear owners, review triggers, and metadata, merchants can control modes, knowledge, schedules, tags, and handoff rules. A compact review card should accompany every change, outlining what changed, where the fact is stored, expected questions, allowed answers, escalation procedures, and the owner. Testing should focus on behavior rather than eloquence, verifying that AI responses use correct facts, preserve important conditions, acknowledge uncertainty when needed, and transfer when judgment or external action is required. Review notes should identify the cause of failures, whether it's missing knowledge, conflicting information, incorrect retrieval, unclear boundaries, broken routing, or weak presentation. Addressing these issues requires source corrections, boundary adjustments, routing corrections, or new regression questions. Finally, merchants must recognize failure modes such as insufficient ownership of knowledge fixes, conflicting maintained sources, silent source selection by AI, undocumented handoffs, and automation running outside intended coverage rules. In these cases, an explicit limitation and a useful transfer to a human should be provided, preserving the original intent, relevant context, already checked facts, and the reason for automation's failure.",
  "summary": "AI support becomes useful only when the operating rules behind it are explicit. The hard part is rarely writing a fluent reply; it is deciding which facts are authoritative, what the assistant may do, and when a person must take over. This article applies that discipline to merchant control . The practical goal is: Keep modes, knowledge, schedules, tags, and handoff rules visible and adjustable…",
  "key_points": [
    "Define authoritative sources, scope, conditions, and handoff procedures for AI support.",
    "Model knowledge as maintained data with clear owners, review triggers, and metadata.",
    "Focus testing on behavior, verifying AI responses use correct facts and acknowledge uncertainty."
  ],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}