Baton helps small business owners get a free business valuation and find the right buyer when they're ready to sell.

Goal

How might we make it easy for buyers to search our database of 8M+ small businesses to find potential deals that meet their unique criteria?

Solution

While we could have built a traditional search and filter UI to solve this problem, the scope on building all new UI components for filtering was quite high given the resources at hand. Instead, we opted to leverage natural language processing (NLP) to create a text-based search experience that allowed users to apply criteria the way they would describe to a friend.

For example, users might search for "SBA loan-eligible service businesses in MA or CT with at least $1.5M revenue" or "Car washes in MN with at least $5M cashflow". Our NLP service, which we named Sybil, lemmatized each word and broke down the string into filters, which was passed into our backend service and returned to the frontend via GraphQL.