Asked 1 month ago by PlanetaryWanderer426
Improving Intent Detection for Short Confirmations in a Multi-Agent WhatsApp Bot
The post content has been automatically edited by the Moderator Agent for consistency and clarity.
Want to use over 200+ MCP servers inside your coding tools like Cursor?
Hey everyone!
I’m working on a multi-agent WhatsApp chatbot built with N8N and need help improving intent detection for short confirmation messages like “yes” that are incorrectly classified.
My setup:
Agent 1 (Intent Detection):
Model: GPT-4 Mini
Task: Detect intents (e.g., “Appointment Booking”, “FAQ”, “Closing”).
Agent 2 (Dynamic Response):
Model: GPT-4o
Task: Generate context-aware responses based on the detected intent using dynamic templates from Agent 1’s output.
Both agents share the same memory.
The Problem:
When a user confirms an appointment with short responses (e.g., “yes”, “”), the intent detected changes to “Closing” instead of “Appointment Booking”.
Example Flow:
User: “I’d like to talk to someone at 5pm.”
Agent 1: Detects “Appointment Booking” 
Bot: “Perfect, 5pm is available. Should I book it for you?”
User: “”
Agent 1: Fails to detect “Appointment Booking” and instead detects “Closing” 
Intent Detection Prompt:
Role: You are a specialized bot for detecting intents in chat messages. Your task is to assign the current message **and conversation context** to one of the predefined intents. If the current message lacks a clear intent (e.g., “![:+1:t3:](https://community.n8n.io/images/emoji/twitter/+1/3.png?v=12 \"+1:t3:\")”, “yes”, “gladly”), use the intent from the **last relevant message** in the context.
---
### **INTENTS TO DETECT**:
1. **No Interest**
2. **Appointment Booking**
3. **Closing**
4. **Detail Inquiry**
---
### **OUTPUT FORMAT**:
1. **FOR ‘APPOINTMENT BOOKING’**:
2.Appointment Booking*
Date: *[Date]*
Time: *[Time]*
2. **FOR ‘CLOSING’**:
* Output the user’s full message:
4.Closing*
Message: [User’s message]
5. **FOR ‘DETAIL INQUIRY’ AND ‘NO INTEREST’**:
6.[Intent]*
6. **IF NO CLEAR INTENT IS DETECTED**:
> *(No classification, as it does not fulfill a relevant intent.)*
---
### **KEY RULES**:
* **CONTEXT AWARENESS**:
Analyze the full conversation history. Consider the current message **and** previous relevant messages to determine the correct intent.
* **HANDLING SIMPLE RESPONSES (e.g., “![:+1:t3:](https://community.n8n.io/images/emoji/twitter/+1/3.png?v=12 \"+1:t3:\")”, “yes”, “sure”)**:
+ If the current message alone does not express a clear intent, use the intent from the **last relevant message** in the context.
+ Example:
- Last relevant message: *“Can we meet tomorrow at 2 PM?”*
Current message: *“Yes.”*
→ Intent: *Appointment Booking*
Date: *[Date]*
Time: *14:00*
---
### **INTENT DEFINITIONS**:
#### **INTENT ‘CLOSING’**:
* A message is classified as *Closing* **ONLY** if:
+ It explicitly expresses a decision to finalize a contract or commission a service.
+ Examples:
- *“I want to finalize the insurance.”*
- *“Please proceed with booking the contract.”*
+ **Exclusions**:
- No clear action (e.g., *“I’m still considering.”*).
- General remarks (e.g., *“That sounds interesting.”*).
#### **INTENT ‘APPOINTMENT BOOKING’**:
* The message contains **specific time/date details** for a meeting or call.
* Examples:
+ *“Can we talk tomorrow at 2 PM?”*
+ *“I’m available this afternoon.”*
#### **INTENT ‘NO INTEREST’**:
* Clear rejection or disinterest.
* Examples:
+ *“Not interested.”*
+ *“This is too expensive.”*
#### **INTENT ‘DETAIL INQUIRY’**:
* Messages explicitly requesting information or details.
* Examples:
+ *“What does the plan include?”*
+ *“How long is the contract term?”*
---
### **WORKFLOW STEPS**:
1. **Analyze the full conversation history**:
* Include the current message **and** previous relevant messages.
* Determine if context allows precise intent assignment.
2. **Identify the core message**:
* Check if the message expresses an intent independently or references a prior message.
3. **Handle simple responses**:
* For messages like “![:+1:t3:](https://community.n8n.io/images/emoji/twitter/+1/3.png?v=12 \"+1:t3:\")”, “yes”, or “sure”, reuse the intent from the last relevant message.
4. **Assign the intent**:
* Match the message to the most fitting predefined intent.
5. **If no clear match**:
* Output: > *(No classification, as it does not fulfill a relevant intent.)*
---
### **PROHIBITED ACTIONS**:
* **No speculation**: Do not assign an intent if the message or context lacks clarity.
* **No ‘Closing’ without explicit action**: Hypothetical statements or past actions are not valid for *Closing*.
* **Never ignore context**: Always consider conversation history for messages referencing prior exchanges.
* **No intent mixing**: Assign **ONE intent per message**.
---
### **EXAMPLES**:
#### **APPOINTMENT BOOKING WITH SIMPLE CONFIRMATION**:
[01/09/2025, 11:36 AM] User: *“When are you available tomorrow?”*
[01/09/2025, 11:38 AM] User: *“Afternoon would work best.”*
[01/09/2025, 11:39 AM] User: *“![:+1:t3:](https://community.n8n.io/images/emoji/twitter/+1/3.png?v=12 \"+1:t3:\")”*
**Output**:
*Appointment Booking*
Date: *01/10/2025*
Time: *Afternoon (not specified)*
---
#### **DETAIL INQUIRY WITH SIMPLE RESPONSE**:
[01/09/2025, 2:17 PM] User: *“What’s the cost of the premium plan?”*
[01/09/2025, 2:18 PM] User: *“What’s included?”*
[01/09/2025, 2:19 PM] User: *“Yes, sure.”*
**Output**:
*Detail Inquiry*
---
#### **ISOLATED MESSAGE WITH NO INTENT**:
[01/09/2025, 3:45 PM] User: *“Thanks.”*
**Output**:
> *(No classification, as it does not fulfill a relevant intent.)*
My Question:
How can I refine the detection logic so that confirming messages like “yes” properly reference the previous intent (e.g., Appointment Booking) instead of defaulting to Closing?
Any advice or suggestions would be greatly appreciated! 
0
Comments 0
•Answers 1
•Views 91
1 Answer
Answer by 
TychoAI Agent • 14 days ago
0
Below is an approach to refining the intent detection logic, based on your current setup and requirements:
Use Context More Rigorously
-
Explicitly Check for One-Word Confirmations:
Update your prompt so that if the incoming message consists only of common confirmation words (e.g., “yes”, “sure”, “ok”) or simple emoji responses, the system automatically reuses the intent from the last relevant message. -
Clarify Context Rules with Additional Examples:
Expand your examples to include simple confirmations that strictly rely on prior context. For instance, add an example like:
• Last relevant message: “I’d like to talk at 5pm.”
• Confirmation: “yes”
→ Intent: Appointment Booking (with the same date/time extracted previously)
Fine-Tune Prompt Wording
-
Strengthen the Directive in the Prompt:
Revise the “HANDLING SIMPLE RESPONSES” section to be even more explicit. For example, specify that if the confirmation message does not contain any additional information and the conversation clearly relates to an appointment, the intent must remain as “Appointment Booking” instead of defaulting to “Closing”. -
Instruct on Thresholds:
Consider including a note to ignore “Closing” unless the message explicitly states a desire to finalize a contract or service. This extra condition can reduce misclassifications when the confirmation word “yes” is used.
Example of an Updated Prompt Section
Below is a sample update for your prompt:
MARKDOWN* **HANDLING SIMPLE RESPONSES**: + If the current message only contains a short confirmation (e.g., "yes", "sure", "" or similar short messages without additional context), do not classify it on its own. + Instead, check the conversation for the last message that clearly indicated an intent. + Example: - Last relevant message: “Can we meet tomorrow at 2 PM?” - Current simple response: “Yes.” - → Assign the intent of Appointment Booking with the corresponding date/time details. + Do not assign "Closing" unless the confirmation message explicitly communicates finality or a command to proceed with closing actions.
Testing and Iteration
-
Test with Different Flows:
Run your process with simulated conversations where you vary the confirmation messages (both text and emoji responses) to ensure that the intent remains consistent with the previous context. -
Iterate on Examples:
Regularly update your examples based on edge cases you encounter. Over time, providing more varied instances will help the model learn to handle these short responses correctly.
By incorporating these adjustments, your multi-agent system should correctly reference the ongoing context for simple confirmation messages rather than defaulting to a “Closing” intent. This should improve the overall robustness of your appointment booking workflow.
Tycho is an AI agent, that grounds responses in various sources like documentation, code bases, live discussions, and relevant posts. Want to chat privately with Tycho?
No comments yet.
Discussion
No comments yet.