AI - suggested tags

AI suggested tag candidates

Relying on strong AI and natural language processing techniques, Relevance AI allows you to use AI to extract tag candidates from your dataset.

This is done under two options:

• high level (i.e. topics -> example traffic, food )
• low level (i.e. deeper analysis -> example delicious food, heavy traffic)

Note: High-level suggestions are normally one-word terms (e.g. activities), whereas low-level suggestions are often composed of a few words (fun outdoor activities).

Generate AI-suggested tags

Under taxonomy builder and from the right-hand side,

• click on "Suggestions"
• select the field whose data is used for extracting candidate tags
• select the granularity level (i.e. high or low level)
• click on "generate ... topics"

Note: This is a complex process and the required time depends on the size of your dataset. For instance, for a dataset of 10K long responses, the required time is estimated to be between 30 to 45 minutes.

Review tag suggestions

After AI processing is done, the generated list of candidate tags will be shown on the right side.

Make sure to refine the suggested tag list using your domain knowledge.

• Clicking on the suggested tags (seen on the right) will add them to the final list (seen on the left)
• On the left side, use the pencil icon to modify the text
• You can type in new tags on the left hand side. Hit enter after each tag
• You can copy paste a comma-separated tag list on the left hand side

Save and use the tags

When the list is finalized hit Save or Save as new on the top right.

Your tag list will be saved and shown in a window similar to the image below. Here, you can

• Change the label/title under which the taxonomy is saved using the pencil icon next to the label
• Create new taxonomy via Create new taxonomy link on the top right
• Modify an existing taxonomy via the Edit link
• Duplicate an existing taxonomy using the duplicate icon
• Delete an existing taxonomy using the bin icon
• Use an existing taxonomy (e.g. load them into the AI-tagging workflow) via Use this taxonomy link