How to prevent SPAM while using AI Assistant and FLOWS?

[agk]

Dear Pabbly Team,

Great work on Pabbly Chatflow.

I have a query. How to prevent someone from spamming whatsapp when AI or FLOWS is enabled? Guidelines regarding the same is appreciated. Someone might misuse the AI leading to higher API costs and Pabbly Credits.

 

[Preeti Paryani]

Hello @agk,

Thank you for your message and for your appreciation of Pabbly Chatflow.

To help avoid spamming and control unnecessary usage of WhatsApp flows when AI or custom flows are enabled, we recommend the following:

1. Use Specific Keywords:
   Avoid using very common or generic keywords like "hi" or "yes" to trigger flows, as they can be easily misused. Instead, use more specific or unique keywords that are less likely to be triggered accidentally.
2. Set a Stop Word for AI Assistant:
   If you're using the AI Assistant, you can define a specific "stop word" or keyword (such as “end” or “stop”) to prevent further responses. This allows you to give users control over the conversation and helps you avoid unnecessary usage caused by spammers.

Implementing these simple strategies can significantly reduce the chances of spamming and help you manage credit/API usage more efficiently.

 

[agk]

Thankyou for your inputs.

1. I have a FLOW which starts with any keyword. It is essential for me as I want the user to automatically get into the flow. Is there a way to rate limit the FLOW itself. Like customer cannot trigger a particular FLOW more than thrice in an hour. Something like that.
2. I do not understand this one. Spammers would not message us "stop" or "end" as they want to spam. How can I control or rate limit the AI assistant? Currently I am using the AI assistant inside FLOWS. It works well as it is a single AI message. However if the FLOW can be triggered unlimited times, then there is a problem.

Rate limit for a FLOW or AI assistant might solve the problem. Please take it as a feedback or if there is something better pl let me know.

 

[Preeti Paryani]

Hello @agk,

Thank you for your input.

Currently, there’s no built-in rate limit on triggering a Flow for a contact. However, you can:

• Use specific keywords to control when a Flow is triggered.
• Train the AI assistant to avoid repeating the same information to the same user.
• Set a stop word (e.g., “human” or “support”) to trigger a fallback message and limit retry attempts to reduce unwanted responses.

We’ve shared your feedback with our product team for future consideration. Let us know if you need help setting any of this up!