# What can you build

You can build custom A.I. models or create a pre-trained model. Pre-trained models are a great way to quickly experience the platform; they cover many common use cases. Using custom models, you can provide your own labels and make your own custom predictions.

Here's a summary of what you can build:

Model Type Training Description
Sentiment analysis pre-trained Get a sentiment prediction for a text input. For example, you can predict whether a review is postive, negative or neutral
Common names and places recognition pre-trained Identify names and places mentioned in a text input. For example, you can pass in the text of a contract and get a list of company and people names mentioned in the contract
Custom names and places recognition (NER) user trained Identify various custom datapoints mentioned in a text input. For example, you can pass in the text of a work order receipt and get the PO numbers listed in the work order
Multi class text classification user trained Get a category for a given text input using your own category labels. For example, you can flag an incoming customer request with a priority level such as P1, P2 or P3
Multi label text classification user trained Get multiple category predictions for a given text input. For example, you can flag an incoming customer request with request topics.