AI | PROJECT BUILD | APPLICATION MODERNIZATION |

Central Reach

Transforming Healthcare Data Retrieval with AI

Download PDF

Summary:

Our client, Central Reach is a leading AI-powered healthcare software provider, specializing in Applied Behavior Analysis (ABA) and multidisciplinary therapies for autism spectrum disorder (ASD) care and faced significant challenges managing complex, unstructured patient data. Their Learning Tree system stored diverse information including demographics, medical history, treatment progress, and associated documents.

However, the system’s lack of a strict schema and reliance on unstructured data made retrieving specific information cumbersome and inefficient. Clinicians required an intuitive way to interact with the data using natural language queries to support evidence-based practice. Recognizing the need for a more efficient solution, the client sought an AI-powered system capable of handling complex, unstructured data and enabling seamless interaction through natural language queries.

Tech Stack:

Database: MongoDB
Backend: FastAPI
AI: LlamaIndex, MongoDB Atlas

Our AI-powered system introduced an intelligent approach to retrieving and managing Learning Tree data. The solution was built around two key components: a Data Ingestion and
Enrichment Pipeline and an Agentic Retrieval-Augmented Generation (RAG) Service using LlamaIndex and MongoDB.

The Data Ingestion and Enrichment Pipeline structured raw Learning Tree data to enhance searchability through taxonomy processing, standardizing data classification using Learning Tree Standard Reference Markdown files. The Learning Tree Analysis extracted hierarchical relationships and metadata, while data enrichment summarized nodes with Large Language Models (LLMs) and categorized them into structured formats.

The high-level approach of the gravity9 solution centers on leveraging advanced AI technologies to enhance data retrieval and response quality for clinicians. We designed an approach that integrates natural language processing (NLP) with database querying to enable seamless interaction with complex Learning Tree data stored in MongoDB. The system architecture consists of two main components: the Data Ingestion and Enrichment pipeline and the Agentic Service, which together transform raw data into a structured knowledge base and provide AI-driven query capabilities.

RELATED INSIGHTS

AI, Blog, Partner

Why gravity9 is Proud to Partner with LlamaIndex

With AI agents and Agentic workflows becoming essential in the market gravity9 is excited to partner with LlamaIndex, a powerful framework that enhances AI-driven automation through intelligent data retrieval and structured indexing, ensuring smarter, more efficient AI solutions for our clients

Find out more

AI, Modernisation, MongoDB

gravity9 - Proud to be a Launch Partner for MongoDB AI Application Program

gravity9 furthers its commitment to accelerating the adoption of gen AI among MongoDB customers through MongoDB's AI Application Program

Find out more

AI, Blog

Harnessing OpenAI Large Language Models and Retrieval Augmented Generation

gravity9's Tomasz Maruszak shows how Large Language Models (LLM) and Retrieval Augmented Generation (RAG) yield efficiency and cost savings.

Find out more

It surfaces all that information and puts it right at your fingertips. You don’t have to hunt and peck, or worry if someone filled out the right…

David Stevens, Head of AI at CentralReach

Talk to our AI team

To find out more about this project, our experience within the healthcare sector or to explore the possibilities of agentic AI for your organization, reach out to our AI team today.

Contact Us

"*" indicates required fields