While this type of search can be very accurate, it’s not easy to write. Only experienced researchers can create something this in-depth, and many less-experienced users struggle.

We think there might be a better way.

When AI meets search

Most conversation online contains bits of information that help us understand the context, even in tweets with less than 140 characters.

Take this tweet:

To create a search that would identify this as a post about the company Apple would be pretty tough. There’s no mention of the iPhone or App Store, no mention of Tim Cook or Steve Jobs. The image isn’t even related.

But there are some hints. The capitalized “A” in Apple, the mention alongside another major company (Tesla), and the word ‘investing’ all indicate to us humans that this is a mention of Apple the company, not the fruit.

We’ve now built a machine-learning algorithm to try and understand those same hints.

A smarter way to search

Using smart machine learning, we’ve been able to train an algorithm that analyzes conversation online for these types of hints. By referencing a corpus of conversation around terms like Apple, our algorithm can accurately predict that this tweet is for Apple the company rather than the fruit or record label.

We’ve released this groundbreaking technology into Search, our new unlimited and fast search engine.

This new technology makes searching simple. Rather than spending hours writing a detailed query, simply start typing and Brandwatch will automatically suggest what you’re looking for and do all that hard work for you.