Our NLP engine is completely horizontal. You can use it with any vertical. For the system to work with your vertical, you have to first give it a training corpus. It should consist of historical user queries with the correct answers assigned to them. The system will learn your vertical’s jargon. ...

As a private company, we don't disclose our clients. But you may see our bots in action. This bot you are using right now is the Answer bot. An example of the Slot filler API usage is our travel agent bot on Facebook Messenger. There are more. ...

There is an Analytics tab in the FAQ Editor. It displays common stats such as visitors, queries, conversations, deflected tickets, etc.

Our FAQ API may be used to automate both inbound and outbound calls. It is up to you, the developer. It is the most valuable for inbound calls. There is much variability in what users say. Our API reduces it to a formal format the computers can understand.

We use supervised learning. For the system to work, you have to first give it a labeled training corpus. It should consist of historical user queries with the correct answers or intents assigned to them. Slots shall be identified and labeled.

Precision and recall depends on product/algorithm and your corpus. For Phraser and FAQ we are typically seeing ~88% precision@1, ~95% p@3 (at 100% recall). For slot filling, ~98% per-label precision. Both are the most accurate in the industry.

All our products are available white labeled. You are welcome to use them under your own brand.

This bot was created by Alterra.ai team , using our proprietary Deep Learning - based NLP Engine.

This bot doesn't know anything about time, weather, as well as other external topics. It can only answer questions about our company and products.

Alterra.ai, the company, was founded in 2013. However, the current product line has been developed in the last couple of years.