AI Agent responde con informacion desactualizada

Tengo un chat bot de IA para facebook messsenger. Implemente el chat bot pero en eso cambiaron unos precios, sin desconectarlo de facebook messenger en entrenamiento pedi el cambio de informacion y en esa misma conversacion de entrenamiento me muestra los datos nuevos. Pero en interaccion con usuarios de facebook aun dice la informacion vieja. ya conecte y desconecte en varias veces el agente pero nunca contesta en fabeook con la informacion nueva

Hi Alicia,

Thanks for reaching out to Jotform Support. Our Spanish Support agents are busy helping other Jotform users right now, so I'll try to help you in English using Google Translate, but you can reply in whichever language you feel comfortable using. Or, if you'd rather have support in Spanish, let us know and we can have them do that. But, it might take a while until they're available again.

Now, coming back to your question, I optimized the form's performance by clearing the cache. Can you try clearing the Messenger cache as well? Since you’ve already tried disconnecting and reconnecting the agent, try it again for the last time and make sure you give it enough time for the updated knowledge base to sync properly with the integration afterward.

Give it a try and let us know if you have any other questions.

Thank you for the support. I followed the instructions and was able to successfully connect the AI agent to the Facebook account, and it started reflecting the information correctly. However, it’s displaying behaviors that weren’t present during training, such as constantly mentioning that it’s an AI. I’ve given it instructions not to mention that, but it still responds with things like:

• Before you go, would you like me to email you this conversation? Just let me know which address to send it to. Since I’m an AI, I can help you with that easily.
• Hello, good morning. How can I assist you regarding colposcopy services? I’m an AI and I’m here to help you.

This is something we don’t want it to do, but I’m no longer sure how to prevent it. We’re considering temporarily disconnecting it because of these types of messages. I am not sure if this issue can be followed in this same ticket or if I should open another one.

Hi Alicia,

To provide you with the most effective suggestions for your AI agent, our team would like to clone your AI agent along with its associated knowledge base. This replication allows us to mirror your unique environment and craft a solution that’s perfectly suited to your needs. We guarantee that your data will be treated with the highest level of security and confidentiality, and will only be used to help enhance your experience. After your issue is fully resolved, the cloned AI agent and its data will be deleted.

Can you confirm that we have your permission to go ahead with this process? Should you have any questions or concerns, feel free to reach out.

After we hear back from you, we’ll have a better idea of what’s going on and how to help.

Yes, I would like to authorize the cloning of the AI agent, knowing that the information will be treated confidentially. My question is: when will this happen? My idea was to delete all the base information and start from scratch, so to speak (do you have any video tutorials or specific documentation on how to train the AI agent?), and connect the agent to a Facebook account I have for testing purposes, without reconnecting it to the real page until the responses show consistently appropriate behavior.

However, given the current situation, perhaps I could leave my agent MAYA as it is and start training a different AI agent with the action plan I mentioned above. I will not proceed until I receive instructions from you regarding what would be the most appropriate course of action. For now, my AI agent is disconnected from Facebook due to the messages I mentioned in my previous response.

Hi Alicia,

I’ll need a bit of time to look into this. I’ll get back to you as soon as I can.

In the meantime, let us know if there’s anything else we can help you with. 

At the moment that would be my only query. I am waiting to know how to proceed. I would probably be exploring training another agent but would leave the current one (MAYA) as is. 

Hi aliciasvv,

There are several features you can customize in your AI Agent Builder that affect how your AI Agent interacts with customers, including the greeting message. You can easily change your Jotform AI Agent's greeting message. Let me show you how:

1. In AI Agent Builder, click on the Gear icon on the right side of the Channel window.
2. Then, in the Standalone Settings menu on the right, go to the Greeting tab.
3. Toggle Off the AI-Generated Greeting Message feature and enter your own message in the Greeting Message section.

To bold a word, phrase, or sentence, add two Asterisks before and after it:

As for how your AI Agent replies, you can customize its behavior to avoid mentioning that it is an AI. You can do this by updating the Agent Personality settings. Check out the screenshots below:

Give it a try and let us know how it goes.