ManyChat-ChatGPT Integration Tutorial
Budget: $10 – $20 USD
the duration should be approximately 10 minutes.
divided into scenes.
I provide a prompt for reference and task description.
0:00 – 1:00
In this video I show you how you can connect a Manychat chatbot with ChatGPT,
so that it automatically responds to your customers.
Here's how it works:
I write a question here…
"How long will it take for the order to arrive?"
and the chatbot answers instantly, using artificial intelligence.
Everything is automatic and it only takes a few minutes to set it up.
Let me show you exactly how it's done.
1:00 – 1:30 → WHAT YOU WILL LEARN
In this video we will do 3 simple things:
1. Create a flow in Manychat
2. Send the user's message to ChatGPT
3. Display the automatic response in the chatbot
You don't need programming.
Everything is done visually, step by step.
1:30 – 3:30 → SETUP MANYCHAT
We enter Manychat and go to Flow Builder.
We create a new flow and add a message.
Here we ask the user to write their question.
We add a User Input and select the Text type.
We save the user's message in a field,
which we call "user_question".
This will be the message we send to ChatGPT.
3:30 – 6:30 → CHATGPT CONNECTION (IMPORTANT)
Now comes the important part.
We add an Action and select External Request.
Here we will send the message to the ChatGPT API.
We set the method to POST
and the URL to the OpenAI API.
We add the necessary headers:
Authorization with API key
and Content-Type application/json.
In the Body, we send the following data:
the GPT model
a system message, where we define how the AI should respond
and the user's message, using the user_question variable.
Basically, we tell the AI:
respond like an online store assistant,
short and clear.
After sending the request,
we need to save the response.
In Response Mapping,
we select the generated text
and save it in a field called "ai_answer".
6:30 – 8:30 → SHOW RESPONSE
After this step,
we add a new message to the flow.
Here we display the AI response using the variable:
ai_answer
Basically, any response that comes from ChatGPT,
will be sent directly to the user.
Now your chatbot can automatically answer questions.
8:30 – 9:30 → REAL TEST
Let's test.
I type a question:
"How can I return a product?"
And we see that the chatbot responds instantly,
with a clear and natural message.
This means that the integration is working correctly.
9:30 – 10:00 → RECAP
In this video you learned:
how to create a flow in Manychat
how to send the message to ChatGPT
and how to display the automatic response.
Now you can use this system for customer support,
sales or automation.
In the next steps, you can also add integration with real orders.
I want a short demo first with something similar if you've done it before.!!!
divided into scenes.
I provide a prompt for reference and task description.
0:00 – 1:00
In this video I show you how you can connect a Manychat chatbot with ChatGPT,
so that it automatically responds to your customers.
Here's how it works:
I write a question here…
"How long will it take for the order to arrive?"
and the chatbot answers instantly, using artificial intelligence.
Everything is automatic and it only takes a few minutes to set it up.
Let me show you exactly how it's done.
1:00 – 1:30 → WHAT YOU WILL LEARN
In this video we will do 3 simple things:
1. Create a flow in Manychat
2. Send the user's message to ChatGPT
3. Display the automatic response in the chatbot
You don't need programming.
Everything is done visually, step by step.
1:30 – 3:30 → SETUP MANYCHAT
We enter Manychat and go to Flow Builder.
We create a new flow and add a message.
Here we ask the user to write their question.
We add a User Input and select the Text type.
We save the user's message in a field,
which we call "user_question".
This will be the message we send to ChatGPT.
3:30 – 6:30 → CHATGPT CONNECTION (IMPORTANT)
Now comes the important part.
We add an Action and select External Request.
Here we will send the message to the ChatGPT API.
We set the method to POST
and the URL to the OpenAI API.
We add the necessary headers:
Authorization with API key
and Content-Type application/json.
In the Body, we send the following data:
the GPT model
a system message, where we define how the AI should respond
and the user's message, using the user_question variable.
Basically, we tell the AI:
respond like an online store assistant,
short and clear.
After sending the request,
we need to save the response.
In Response Mapping,
we select the generated text
and save it in a field called "ai_answer".
6:30 – 8:30 → SHOW RESPONSE
After this step,
we add a new message to the flow.
Here we display the AI response using the variable:
ai_answer
Basically, any response that comes from ChatGPT,
will be sent directly to the user.
Now your chatbot can automatically answer questions.
8:30 – 9:30 → REAL TEST
Let's test.
I type a question:
"How can I return a product?"
And we see that the chatbot responds instantly,
with a clear and natural message.
This means that the integration is working correctly.
9:30 – 10:00 → RECAP
In this video you learned:
how to create a flow in Manychat
how to send the message to ChatGPT
and how to display the automatic response.
Now you can use this system for customer support,
sales or automation.
In the next steps, you can also add integration with real orders.
I want a short demo first with something similar if you've done it before.!!!