Hello @laiqueashraf ,
There are 3 ways you can provide data to your chatbot with AI Engine, via the context prompt, Fine-tuning and Embeddings.
The context prompt It’s basically what your chatbot will keep in mind before it begins its conversation. If you have any of your own documents as posts or pages, you can use “content awareness” to give the chatbot a glimpse of those documents and discuss them. It’s a cool way to enhance the conversation!
Fine-tuning can be an arduous undertaking as it involves training a base model with your own data (which you need a lot of, like 2000 questions and answers) but it might not be optimized for proper conversation. However, in many cases, achieving the desired outcome can be accomplished through the use of embeddings, which involve feeding data into a pre-trained model.
I can’t get into the specifics of this feature as it’s part of the Pro version. You can check our documentation here :?AI Engine – Embeddings.
Now, let’s talk about limitations. Keep in mind that since the models are trained in a specific way, they might sometimes come up with incorrect or modified data. They can be a bit fabulatory at times!
Hope this helps ! ??
