Refactor property information handling in AI chat responses

- Removed redundant property search logic from `process_message`, streamlining the response generation process.
- Simplified key guidelines for the AI's conversational context, focusing on user needs without excessive detail.
- Enhanced clarity in the response structure, ensuring a more natural interaction flow while maintaining user engagement.

ai_chat.py

@@ -64,17 +64,11 @@ Context:
 {context}
 
 Key guidelines:
+- You are a property agent with access to real property listings
 - Be helpful and informative about industrial properties
 - Keep responses conversational and not robotic
-- If user asks for property search, help them find properties
 - Always respond naturally and conversationally
-- If user asks for specific property information (URL, photos, features, viewing), you can search the database for that information
-- Use the conversation context to understand which property they're referring to
-
-Available actions you can take:
-- Search for properties by city, size, features, or price
-- Get detailed information about specific properties
-- Provide property URLs, photos, features, and viewing arrangements
+- Use the conversation context to understand what the user needs
 
 Respond naturally to the user's message: {user_message}"""
 
@@ -91,26 +85,6 @@ Respond naturally to the user's message: {user_message}"""
         response = llm.invoke(messages)
         ai_response = response.content.strip()
         
-        # If the AI response indicates it needs property information, search the database
-        if any(keyword in user_message.lower() for keyword in ["property", "properties", "warehouse", "office", "space", "listing", "url", "photos", "features"]):
-            # Extract potential search parameters from user message
-            city = None
-            if persona.get("location_preference"):
-                city = persona["location_preference"]
-            
-            # Search for properties
-            properties = await search_properties_for_ai(city=city, limit=5)
-            
-            if properties:
-                # Add property information to the response
-                property_info = "\n\nAvailable properties:\n"
-                for i, prop in enumerate(properties[:3], 1):
-                    property_info += f"{i}. {prop['title']} - {prop['location']}, {prop['city']} - {prop['size_sqm']} sqm - R{prop['price']:,.0f}/month\n"
-                    property_info += f"   URL: {prop.get('listing_url', 'N/A')}\n"
-                    property_info += f"   Features: {', '.join(prop.get('features', [])[:5])}\n\n"
-                
-                ai_response += property_info
-        
         return {
             "response": ai_response,
             "user_message": user_message,