We are developing a chatbot related to device troubleshooting which is following a flow based on confirm or deny intents and different questions are asked by bot based on the confirm or deny intents for different devices.
I trained my RASA Core model for 2 devices with all possible different confirm or deny flow stories with utter messages defined in yaml file for both devices.
Now when I started testing core model, if I started for a device, after confirm or deny intent, it started returning utter message of second device.
I don’t understand fully how you differentiate your stories and data for the two devices.
Can you maybe post a few short examples from your domain file, stories and nlu data?
Do you use the same intents for both devices?
Depending on how similar the devices and conversations are an option would be to train separate models for each device, this way you can reuse your nlu data and stories and only change the utter templates in the domain file in order to be device specific and the model cannot predict actions for a different device.
Please let me know if this helps and if you have further questions.