The WhatsApp broadcast that started talking to itself: Eliminating auto-responder loops

The Absence of `Auto-Submitted` Headers on WhatsApp

In conventional messaging protocols, machine-generated traffic carries explicit structural headers:

• SMTP Email: RFC 3834 Auto-Submitted: auto-replied, Precedence: bulk.
• Slack API: bot_id parameter present on message payload.
• Telegram Bot API: User.is_bot: true.

The Meta WhatsApp Cloud API provides no such header. An inbound webhook payload delivers identical fields (from, wa_id, text) whether written by a human typing on a keyboard or generated by an Android out-of-office auto-reply app:

{
  "object": "whatsapp_business_account",
  "entry": [{
    "changes": [{
      "value": {
        "messaging_product": "whatsapp",
        "contacts": [{ "profile": { "name": "Customer" }, "wa_id": "34600000001" }],
        "messages": [{
          "from": "34600000001",
          "id": "wamid.HBgL...",
          "timestamp": "1717584000",
          "text": { "body": "Thank you for reaching out! We are currently unavailable." },
          "type": "text"
        }]
      }
    }]
  }]
}

const AUTOREPLY_PATTERNS = [
  /unknown\s+input\s+type\s+received/i,
  /kindly\s+reply\s+with/i,
  /thank\s+you\s+for\s+(reaching\s+out|contacting|messaging)/i,
  /gracias\s+por\s+(escribirnos|contactarnos|tu\s+mensaje)/i,
  /shukriya|dhanyawad/i,
  /we\s*['’]?\s*re\s+unavailable/i,
  /out\s+of\s+office/i,
  /fuera\s+de\s+oficina/i,
  /auto-reply|auto\s+reply|automated\s-response/i,
];

export function isAutoResponderPattern(text) {
  const norm = String(text || '').trim();
  if (!norm) return false;
  return AUTOREPLY_PATTERNS.some((pattern) => pattern.test(norm));
}

Broadcast Attribution & Interactive Immunity

A major trap in WhatsApp CRM architecture is treating standard 1-on-1 operational templates (OTPs, appointment reminders) as mass broadcasts. Klaros inspects origin_campaign_id and is_broadcast flags on the outbound record before applying campaign-level threshold rules.

Crucially, interactive button taps and WhatsApp Flow completions take unconditional priority (isInteractive = true). No matter how fast an interactive payload arrives, a button press is treated as HUMAN 100% of the time.

Generative LLM Prompt Refusal Guards

Even if an automated text bypasses regex matching, generative AI models must be restrained from entering conversational loops. Klaros injects an explicit AUTO-RESPONDER GUARD into the model context when broadcast metadata is present:

const AUTO_RESPONDER_GUARD = `
AUTO-RESPONDER GUARD:
If the user's message appears to be an out-of-office auto-reply, automated acknowledgment, or system error (e.g., "Thank you for messaging us", "We are currently unavailable", "Unknown input type"):
1. Set "canAnswer": false
2. Set "handoff": false
3. Do NOT generate a conversational reply.
`;

Competitive Analysis & Technical Benchmark

Empirical Verification & Test Proofs

Every assertion in this note is backed by automated test suites running against real SQLite harnesses and Cloudflare D1 mocks:

# Run broadcast autoresponder loop triage test suite
npx mocha test/uni-cloud-broadcast-autoresponder.test.mjs

# Run correspondent self-match verification
npx mocha test/uni-cloud-correspondent-self-match.test.mjs

# Run full project test suite (9,411 passing tests)
npm test