Embeddings

2. Vector Embeddings

Embeddings allow for improvements in search. Almost every company can implement vectors even if the improvement is only 5-10% for large companies this adds up. For instance a company has a catalogue of 100,000 products.

• A user calls asking for a specific part but doesn't know its name, it describes the product and due to the embeddings gets a better search result leading to faster and more sales.
• A lawyer wants to find specific case law for a niche case. The embeddings allow searching through all past case law more efficiently and in a new way than traditional search models.

100 billion$