The Tagging Validation stage begins once you have been granted access to your project which contains a sample of your data tagged according to your new theme structure.

This exercise is your opportunity to make any significant changes to your tagging, before we deploy your model.

Please go through each of the categories/themes and ensure you are happy with how the theme tags have been applied.

The sample is large, so you do not need to check every single comment.

Provide all your feedback on the document provided - your Implementation Manager will cover this during your Tagging Review session.

To ensure alignment, we recommend that only the stakeholders who participated in the Theme Structure Review also participate in the Tagging Review.

Incorrect sentiment - you feel a theme tag has been correctly applied, but the sentiment is incorrect.

Remove theme tag - you feel a theme tag has been unnecessarily or incorrectly applied to a comment.

Add theme tag - you feel a theme tag is missing from a comment and needs to be added.

Model deployment is when we activate the algorithm trained on the comment sample you reviewed. From this point forward, the model will process the remaining backdate and all new, incoming data.

This is why the validation exercise is your opportunity to make requests significant changes to your tagging.

Once this validation exercise is complete, your model will be deployed. 🎉

No, the number of theme tags is not limited and will be based on the comment context and content.

Yes, only full Admins will be able to change the tagging on the individual comment level. This includes adding themes, removing themes and amending sentiment.

We recommend doing this only on an ad-hoc basis.

If you feel that the theme tagging/and or sentiment for a larger segment of comments needs amending, this would require model retraining. Reach out to your CSM who will liaise with the Machine Learning team to conduct a review.