Introducing Pretectum Cognito Search


Pretectum CMDM offers a sophisticated search experience with "Pretectum Cognito Search," which integrates large language models (LLMs) and Elasticsearch to provide intuitive data retrieval.

Here’s a breakdown of its key features:

Triple Combination Search: This allows users to initiate a search by simply entering a string. Under the hood, it leverages Elasticsearch for initial matches.

LLM-Constructed Query Builder: This is the core innovation of Pretectum Cognito Search. When a simple string search might not yield optimal results, especially when users aren’t familiar with specific field names or tags, the LLM-constructed query builder steps in.

It uses its knowledge of the application’s schemas and tags to interpret natural language questions.

It then constructs a more precise query based on this understanding, aiming to deliver better results.

This eliminates the need for users to learn complex query syntax, making data access more accessible.

Interactive Query Builder: If users receive results from the LLM-constructed query and wish to refine it further without posing another natural language question, they can seamlessly switch to an interactive query builder. This builder still doesn’t require knowledge of query syntax, offering a user-friendly way to adjust the search criteria.

Essentially, Pretectum Cognito Search aims to bridge the gap between user intent (expressed in natural language) and the technical complexity of data querying, by leveraging AI to facilitate more accurate and efficient searches within the Pretectum CMDM platform. This is particularly beneficial for achieving a "Single Customer View" by consolidating and making customer data easily searchable.