Forced to re-train and publish in each instance

[Kim Sullivan]

We have fully built out our Virtual Agent with NLU.  However, the training and publishing are not being captured in an update set.  We have to manually re-train in each instance.

Anyone else run into this?

[User317960]

Kim / Willem,

Adding this here for visibility as well, and for other folks in the future.

You should be able to follow the below steps.

1. Go to sys_nlu_model.list and copy the model name you are trying to add to your update set.
2. Then go to ml_capability_definition_base.list, and use the model name from the first step to search for the solution.
3. From here you can add the model and the solution definitions to the current update set.