CONCEPT SEARCH

How It Works

Content Analyst Company is the original patent-holder for LSI or Latent Semantic Indexing, a mathematical way to index and search for text. This, along with numerous other patents, gives CAAT a novel way to identify related or relevant documents, words, and text based on conceptual similarity.

Unlike other advanced search techniques, CAAT trains itself from material being indexed, uncovers conceptual similarity to an nth-order degree, and does this without the need for dictionaries, word lists, or thesauruses. CAAT’s algorithms have been carefully programmed to be repeatable, so there is no “black box” magic to CAAT and concept search – just incredible searching power to find truly relevant documents.

How it is Used in eDiscovery

Concept Search goes to findability. With keyword searches, it is always trial and error: did I put in all the right words, how can I exclude extraneous and irrelevant information, how many times do I need to refine a search?

Concept Search is very different. The complaint document that triggers a legal dispute can be entered as a search string in its entirety - and CAAT will find every document across hundreds of thousands or even millions of other documents that deal with similar issues. Unlike keyword-only searching, with CAAT’s Concept Search engine, more is better.

People communicate with one another conversationally. CAAT’s concept search engine is very similar to that paradigm: the better a searcher can describe the concept for which they are searching, the high the degree of accuracy of relevant documents CAAT will return. Entire paragraphs are often used as search strings.

CAAT’s powerful “find similar” capability allows searchers to simply select a document and CAAT will find all other relevant documents that are conceptually similar. Using one document to find others is faster than other search techniques. Companies who have implemented the CAAT search engine into their review platforms have seen a significant improvement in how quickly attorneys and paralegals can review documents. Why? Because when someone is looking at similar documents, they quickly adapt to the nomenclature of that given concept and become faster and more accurate reviewers. If for example, they’re looking at documents about banking transactions, then documents that also talk about wire transfers or lock-box transactions are easily understood in the same context. In the case of linear reviews, the reviewer is stuck with a single custodian’s documents – or a specific date range – and those documents will cover a wide range of unrelated topics, slowing review down to a crawl, and accuracy suffers.