Stuck with Flowise post-processing: Formatting or Logging – Not Both

[natitkachenko-max]

Hi everyone,
I’m currently learning AI integrations and could really use your help with a problem I’ve run into.

I’ve built a custom AI RAG assistant workflow using Flowise. I log every user question and answer to Google Sheets using a Make.com webhook via post-processing in Flowise.

The issue is this:
I can't get both proper answer formatting in the widget and accurate, consistent logging to Google Sheets to work at the same time.

Here's what happens:

• If I log the variable Answer (i.e., rawOutput) in the webhook payload, the formatting is preserved, but some answers differ from what is actually shown to the user in the widget.
• If I log the variable answer (modified inside the post-processing script), the logged answer matches what the model says in the widget, but the formatting breaks completely (extra quotes, backslashes, flattened text, etc.).

I tried declaring let answer = Answer to preserve formatting while keeping things in sync — but as soon as I do that, the formatting breaks again.

So I’m stuck: either I get clean formatting OR consistent logging, but not both.

I know I’m missing something obvious, but I can’t figure it out.

Would really appreciate any help, suggestions, or workarounds!

Thanks in advance 🙏
(I’ll add the full code snippet below)

`// === CONFIG ===
const webhookUrl = 'https://hook.us2.make.com/5tnxk82sblhsf7p14xrocdl273w60q26';

// === MAIN FUNCTION ===
async function main() {
  const question = $flow.input || "";
  const Answer   = $flow.rawOutput || "";

  const qLower = question.toLowerCase();
  const aLower = Answer.toLowerCase();

  // --- SPAM FILTER ---
  const ecommerceKeywords = [
    'order', 'shipping', 'delivery', 'return', 'refund',
    'size', 'fit', 'exchange', 'cancel', 'address',
    'payment', 'card', 'visa', 'mastercard', 'paypal',
    'discount', 'sale', 'coupon', 'loyalty', 'points',
    'product', 'item', 'stock', 'restock', 'availability',
    'sweater', 'wool', 'jeans', 'underwear', 'swimwear',
    'clothes', 'store', 'shop', 'purchase', 'policy'
  ];

  const offTopicPatterns = [
    'tell me a joke', 'joke', 'poem', 'write a poem', 'what is love',
    'who is the president', 'capital of', 'weather', 'translate',
    'quantum', 'physics', 'ielts', 'english exam', 'chatgpt', 'openai'
  ];

  const hasShopWord = ecommerceKeywords.some(k => qLower.includes(k));
  const isOffTopic  = offTopicPatterns.some(k => qLower.includes(k));
  const isVeryShort = question.trim().length < 6;

  const isSpam = isOffTopic || (!hasShopWord && isVeryShort);

  // --- FALLBACK FILTER ---
  const fallbackPatterns = [
    "hmm, i'm not sure", "i'm not sure", 
    "i don’t have enough information", "i don't have enough information", 
    "i do not have enough information", 
    "i don’t have information about", "i don't have information about",
    "i’ve forwarded your question", "i've forwarded your question", 
    "i don’t know", "i don't know",
    "i'm unable to make calls", "i cannot connect you",
    "please contact customer support", "please contact"
  ];

  const isFallback = fallbackPatterns.some(p => aLower.includes(p));

  // --- MANUAL ESCALATION (USER REQUESTED HUMAN) ---
  const manualEscalationTriggers = [
    "human", "agent", "representative", "operator", "manager",
    "real person", "talk to someone", "talk to a person",
    "someone help", "need help", "speak to",
    "call an agent", "support agent", "speak with"
  ];
  const wantsHuman = manualEscalationTriggers.some(p => qLower.includes(p));

  // --- ESCALATION / RAG SUCCESS ---
  let ragSuccess = "Yes";
  let escalated = "No";
  let escalationReason = "";

  if (isSpam) {
    ragSuccess = "No";
    escalationReason = "spam_or_off_topic";

    answer =
      "I’m your shopping assistant and can only help with questions about our store — products, orders, shipping, returns, and payments. Please ask me something related to the shop.";
  }

  else if (isFallback) {
    ragSuccess = "No";
    escalated = "Yes";
    escalationReason = "fallback_not_sure";

    answer =
      "I don’t have enough information to answer this one. I’ve forwarded your question to a human support agent — they will get back to you shortly.";
  }

  else if (wantsHuman) {
    ragSuccess = "Yes"; // answer itself was valid, but user requested escalation
    escalated = "Yes";
    escalationReason = "user_requested_human";

    answer =
      "No problem — I’ve forwarded your message to a human support agent. They’ll reply shortly.";
  }

  // --- PAYLOAD ---
  const payload = {
    ts: new Date().toISOString(),
    session_id: $flow.sessionId || "",
    chat_id:    $flow.chatId || "",
    channel:    "website_widget",
    question:   question,
    answer:     Answer,
    rag_success: ragSuccess,
    escalated:  escalated,
    escalation_reason: escalationReason,
    thumbs:     "",
    comment:    ""
  };

  // --- WEBHOOK ---
  try {
    await fetch(webhookUrl, {
      method: 'POST',
      headers: { 'Content-Type': 'application/json' },
      body: JSON.stringify(payload)
    });
  } catch (err) {
    console.error('Webhook error:', err);
  }

  return answer;
}

return main();
`