# US 2026/0229326 A1 · Ai orchestration system for health data analysis

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2026/0229326 A1
- Document ID: US20260229326A1
- Status: Published application
- Kind: A1
- Application number: 19/417,745
- Filed: 2025-12-12
- Published: 2026-08-06
- Earliest claimed: 2025-02-03
- Assignees: GeoOrchestrationAI Corporation
- Applicants: GeoOrchestrationAI Corporation
- Inventors: Ajay Kumar Gupta; Ramani Peruvemba; Paul A Churchyard
- Web page: https://patentgrep.com/patent/US20260229326A1

## CPC

- G16H 10/60: for patient-specific data, e.g. for electronic patient records
- G06F 40/30: Semantic analysis
- G06F 40/56: Natural language generation
- G16H 50/70: for mining of medical data, e.g. analysing previous cases of other patients
- G06F 40/35: Discourse or dialogue representation
- G16H: HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

The present disclosure provides a computer-implemented method for processing health data queries. The method includes receiving a user query through a conversational interface, creating vector embeddings of the user query, comparing the vector embeddings against embeddings of available health metrics stored in a vector database to identify relevant health metrics, extracting geographical parameters from the user query using fine-tuned large language models, retrieving health metric data from a cloud data lake based on the identified relevant health metrics and geographical parameters, filtering the retrieved health metric data based on the geographical parameters, generating summary statistics for the filtered health metric data, and generating multiple output formats including a natural language response and data visualizations.

## Claims

### Claim 1

A computer-implemented method for processing health data queries, comprising:
receiving a user query through a conversational interface;
creating vector embeddings of the user query;
comparing the vector embeddings against embeddings of available health metrics stored in a vector database to identify relevant health metrics;
extracting geographical parameters from the user query using fine-tuned large language models;
retrieving health metric data from a cloud data lake based on the identified relevant health metrics and geographical parameters;
filtering the retrieved health metric data based on the geographical parameters;
generating summary statistics for the filtered health metric data; and
generating multiple output formats including a natural language response and data visualizations.

### Claim 2 (depends on 1)

The method of claim 1, wherein the fine-tuned large language models comprise specialized functions for at least one of selecting metric extent, selecting relevant metrics, selecting geometry levels, checking additional parameters, selecting relevant geometries, selecting administrative filters, selecting administrative filter values, and selecting geometry names.

### Claim 3 (depends on 2)

The method of claim 2, wherein the specialized functions utilize semantic similarity matching algorithms to compare query embeddings with available option embeddings.

### Claim 4 (depends on 1)

The method of claim 1, further comprising initiating cloud-based data processing pipelines when health metric data is not immediately available in the cloud data lake.

### Claim 5 (depends on 4)

The method of claim 4, wherein the cloud-based data processing pipelines include lambda functions that execute automated data processing workflows triggered by events from an event bus.

### Claim 6 (depends on 1)

The method of claim 1, further comprising retrieving geospatial data layers corresponding to the identified health metrics from the cloud data lake.

### Claim 7 (depends on 6)

The method of claim 6, further comprising performing intersection overlay geospatial operations when the geospatial data comprises polygon geometries.

### Claim 8 (depends on 1)

The method of claim 1, wherein the data visualizations comprise maps, tables, pie charts, and bar charts.

### Claim 9 (depends on 1)

The method of claim 1, further comprising storing chat history in a database and updating user queries based on previous interactions.

### Claim 10 (depends on 1)

The method of claim 1, wherein the natural language response is generated using retrieval augmented generation from curated authoritative sources.

### Claim 11

An AI orchestration system for health data analysis, comprising:
a conversational interface configured to receive user queries;
a vector database storing embeddings of available health metrics;
fine-tuned large language models configured to extract parameters from user queries, wherein the fine-tuned large language models comprise picker functions that select relevant options from available datasets based on user queries;
a cloud data lake storing health metric datasets and geospatial data;
a data orchestrator configured to retrieve and filter data based on extracted parameters; and
a visualization engine configured to generate multiple output formats including natural language responses and data visualizations.

### Claim 12 (depends on 11)

The system of claim 11, wherein the picker functions comprise a function to identify geographical areas, a function to identify health metrics, and a function to identify geographic levels.

### Claim 13 (depends on 12)

The system of claim 12, wherein the selectMetricExtent function analyzes user queries to recognize country names, abbreviations, and geographic identifiers for accurate country selection.

### Claim 14 (depends on 11)

The system of claim 11, further comprising cloud data pipeline components including an event bus and lambda functions for processing data requests when health metric data is not immediately available.

### Claim 15 (depends on 14)

The system of claim 14, wherein the lambda functions execute automated data processing workflows that generate required datasets through computational operations including data synthesis and statistical modeling.

### Claim 16

A computer-implemented method for orchestrating health data analysis using artificial intelligence, comprising:
receiving a natural language query from a user;
processing the query using a plurality of fine-tuned large language model functions to extract query parameters, wherein the plurality of fine-tuned large language model functions comprise a function to identify geographical areas, a function to identify health metrics, and a function to identify geographic levels;
retrieving health data from a distributed cloud storage system based on the extracted query parameters;
performing geospatial filtering operations on the retrieved health data; and
generating a comprehensive response comprising natural language explanations and geospatial visualizations based on the filtered health data.

### Claim 17 (depends on 16)

The method of claim 16, wherein the plurality of fine-tuned large language model functions further comprise a function to identify geospatial features and a function to identify administrative boundary levels for data filtering operations.

### Claim 18 (depends on 17)

The method of claim 17, wherein the function to identify geospatial features analyzes user query content to match references against available geometry names within geospatial datasets including hurricane impact areas, drought severity zones, and healthcare facility locations.

### Claim 19 (depends on 16)

The method of claim 16, wherein the geospatial filtering operations comprise performing intersection overlay operations when geospatial data comprises polygon geometries to spatially filter health data points that fall within selected geographic boundaries.

### Claim 20 (depends on 19)

The method of claim 19, wherein the intersection overlay operations utilize computational geometry algorithms that account for coordinate system transformations and boundary precision to ensure accurate spatial filtering results.
