This is where your unlabeled data resides. You can use HumanFirst to explore, organize and annotate your unlabeled data using the following capabilities:

Full-Text Search#

Search by keyword(s) and you will get an updated list of data containing your search query.

Semantic Search#

Semantic search allows you to surface semantically similar utterances from your unlabeled data.


Clustering helps group semantically similar utterances together. This can be helpful when trying to gauge the importance of a topic, or as a tool to quickly select large amounts of variations.

Explore by Similarity#

Explore by similarity helps surface unlabeled data that is similar to the selected intent.

Active Learning Sampling#

When using HumanFirst NLU as the active NLU engine, you can filter your unlabeled data by Margin Score, Uncertainty and Entropy.


The data section also provides basic filters to help explore your data based on various criteria.


The stash is where selected data can be interacted with.