{
  "@context": "https://schema.org",
  "@type": "FAQPage",
  "name": "What should happen when an AI support bot cannot answer a customer?",
  "description": "The workflow should stop guessing and hand the case to a person with the full conversation, customer history, issue category, missing information, and suggested",
  "url": "https://runmyaibusiness.com/faq/what-should-happen-when-an-ai-support-bot-cannot-answer-a-customer",
  "mainEntity": [
    {
      "@type": "Question",
      "name": "What should happen when an AI support bot cannot answer a customer?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Set escalation rules before launch, including what the AI must recognize, what information it must pass along, and who receives the handoff. A useful handoff contains the original conversation, customer history, detected issue, sources consulted, confidence concerns, and a suggested next step. Context: Community examples emphasize escalation with context rather than simply transferring an unanswered message. That prevents the customer from repeating the problem and gives the reviewer a usable starting point. Key factors: - urgency - customer sentiment - policy exceptions - missing or conflicting data - channel response hours. Practical guidance: Test handoffs with angry, vague, multi-part, and off-topic messages. Add a fallback for system outages so customers can still reach a person or receive a clear delay notice. Review failed escalations as carefully as incorrect answers.",
        "description": "The workflow should stop guessing and hand the case to a person with the full conversation, customer history, issue category, missing information, and suggested next step. Test vague, angry, multi-part, and outage-related messages before launch, and provide a clear fallback channel."
      }
    }
  ],
  "speakable": {
    "@type": "SpeakableSpecification",
    "cssSelector": [
      "h1",
      ".faq-question",
      ".faq-answer",
      "[itemprop=\"acceptedAnswer\"]"
    ]
  }
}