How an AI System Answers Financial Questions Using Specialized Bots
This patent describes an AI system where a large language model (LLM) directs specialized machine learning agents to answer natural language questions about financial data, then refines the answers using an adversarial AI.
Original patent title: “Systems and methods of large language model driven orchestration of task-specific machine learning software agents”
This patent describes an AI system where a large language model (LLM) directs specialized machine learning agents to answer natural language questions about financial data, then refines the answers using an adversarial AI. Granted to Broadridge Financial Solutions in 2025 with 23 claims, and it is expected to expire in 2044.
Key facts
Coverage
What does this patent actually cover?
This system takes a user's natural language question about financial records, like "Find similar stocks to Apple." A central large language model (LLM) then processes this request and outputs specific instructions to various machine learning (ML) agents (ClaimclaimA numbered sentence at the end of a patent that legally defines what the inventor owns. The most important section.Read more → 1). For example, a 'similarity-based data record identification machine learning model' (Claim 1) might be instructed to compare financial instruments. These agents perform their tasks, like finding similar financial instruments based on specific attributes, and return their findings. The LLM then takes these findings and generates a natural language response. Crucially, an 'adversarial machine learning agent' (Claim 1) then automatically reviews and modifies this response until it meets certain quality standards, ensuring clarity and accuracy before it's shown to the user.
The gap
What does this patent NOT cover?
- Does not cover systems that process natural language queries for non-financial data records, as ClaimclaimA numbered sentence at the end of a patent that legally defines what the inventor owns. The most important section.Read more → 1 specifies 'at least one attribute associated with at least one financial instrument'.
- Does not cover systems where a large language model directly answers the query without orchestrating separate, task-specific machine learning agents.
- Does not cover systems that identify similar data records without using a 'similarity-based data record identification machine learning model' to determine a metric between items (ClaimclaimA numbered sentence at the end of a patent that legally defines what the inventor owns. The most important section.Read more → 1).
- Does not cover systems that generate natural language responses without an 'adversarial machine learning agent' to recursively auto-correct the output based on predetermined criteria (ClaimclaimA numbered sentence at the end of a patent that legally defines what the inventor owns. The most important section.Read more → 1).
- Does not cover systems where the ML agents are not configured to be instantiated in parallel (ClaimclaimA numbered sentence at the end of a patent that legally defines what the inventor owns. The most important section.Read more → 2).
- Does not cover systems that do not generate an 'explainability prompt' to get a natural language explanation from the LLM about its decisions (ClaimclaimA numbered sentence at the end of a patent that legally defines what the inventor owns. The most important section.Read more → 3).
These exclusions are unique to PatentBrief — derived from the actual claim language, not patent-office boilerplate.
What made this novel
The noveltynoveltyThe requirement that an invention be different from anything publicly known before its priority date.Read more → lies in using a large language model not just to answer questions, but to act as an orchestrator, directing specialized machine learning agents to perform specific data tasks. Even more clever is the 'adversarial machine learning agent' that automatically reviews and refines the LLM's final natural language output, ensuring it meets quality standards before being presented to the user.
The Patent Drawing
Schematic visualization of the patent's claim structure. Hand-drawn diagrams in progress for each landmark patent.
Where you've seen this
Real-world examples
Financial analysis platforms using AI to answer investor questions
Customer service chatbots for banking and investment firms
AI-powered tools for compliance and risk management in finance
Automated financial report generation systems
Intelligent search engines for financial data records
Why it matters
The bigger picture
This patent describes a method for AI systems to handle complex, domain-specific queries, particularly in finance, by combining the broad understanding of an LLM with the precision of specialized ML agents. This approach can improve the accuracy and reliability of AI-generated responses in critical fields. The inclusion of an adversarial agent for self-correction aims to address common issues of factual errors or 'hallucinations' in LLM outputs, making AI more trustworthy for professional use.
Filed
August 1, 2024
Granted
May 20, 2025
Market context
Who's building on this
Companies in this space
Broadridge Financial Solutions Inc., the assigneeassigneeThe entity that owns the patent — usually the inventor's employer or a company.Read more →, is actively developing solutions in this area, focusing on financial technology and data processing. Other major financial technology companies and AI solution providers are also investing in similar architectures to enhance their data analysis and client interaction capabilities, aiming to deploy more reliable and accurate AI for complex financial tasks.
Market impact
This type of technology aims to improve the accuracy and trustworthiness of AI in financial services, potentially reducing operational risks and increasing efficiency. By combining LLM orchestration with specialized agents and self-correction, it could lead to more robust AI tools for financial analysis, compliance, and customer support. This could enable financial institutions to automate more complex tasks, freeing up human experts for higher-value activities and potentially setting new standards for AI reliability in regulated industries.
Claim 1 — Plain English
What this patent covers
This system takes a user's natural language question about financial records, like "Find similar stocks to Apple." A central large language model (LLM) then processes this request and outputs specific instructions to various machine learning (ML) agents (Claim 1). For example, a 'similarity-based data record identification machine learning model' (Claim 1) might be instructed to compare financial instruments. These agents perform their tasks, like finding similar financial instruments based on specific attributes, and return their findings. The LLM then takes these findings and generates a natural language response. Crucially, an 'adversarial machine learning agent' (Claim 1) then automatically reviews and modifies this response until it meets certain quality standards, ensuring clarity and accuracy before it's shown to the user.
The clever bit
The novelty lies in using a large language model not just to answer questions, but to act as an orchestrator, directing specialized machine learning agents to perform specific data tasks. Even more clever is the 'adversarial machine learning agent' that automatically reviews and refines the LLM's final natural language output, ensuring it meets quality standards before being presented to the user.
What it does not cover
- Does not cover systems that process natural language queries for non-financial data records, as Claim 1 specifies 'at least one attribute associated with at least one financial instrument'.
- Does not cover systems where a large language model directly answers the query without orchestrating separate, task-specific machine learning agents.
- Does not cover systems that identify similar data records without using a 'similarity-based data record identification machine learning model' to determine a metric between items (Claim 1).
- Does not cover systems that generate natural language responses without an 'adversarial machine learning agent' to recursively auto-correct the output based on predetermined criteria (Claim 1).
- Does not cover systems where the ML agents are not configured to be instantiated in parallel (Claim 2).
- Does not cover systems that do not generate an 'explainability prompt' to get a natural language explanation from the LLM about its decisions (Claim 3).
Patent timeline
Application submitted to the patent office
Application published, typically 18 months after filing
Patent officially issued
Patent enters public domain
PatentBrief Score
Impact Score
Moderate
Citation count
0/40
No citations yet
Claim breadth
15/20
Broad claimsclaimsThe numbered statements at the end of a patent that legally define what the inventor owns.Read more →
Recency
20/20
Granted within 5 years
Assignee scale
20/20
Major company or institution
PatentBrief Impact Score — based on citation count, claim breadth, recency, and assignee scale. Not a legal assessment.
Heuristic Value Estimate
What this patent might be worth
$31K – $100K
Midpoint $62K · 18.1 yr remaining · industry ×1.6
Heuristic only — blends forward/backward citation counts, claim scope, time remaining, litigation history, and CPC-derived industry baseline. Real valuations need a professional appraisal.
The original legal language
Original claims
23 claims as filed with the patent office.
Concepts involved
Citations
Patent lineage
Cite this patent
Lo, J., Kryeziu, F., & Kwiatkowski, J. (2025). How an AI System Answers Financial Questions Using Specialized Bots (U.S. Patent No. 12,307,349). U.S. Patent and Trademark Office. https://patentbrief.org/patent/us/12307349/systems-and-methods-of-large-language-model-driven-orchestration-of-task-specifi
Auto-generated from the patent record. Double-check author order and the issue date against the official USPTO document before submitting.
Embed
Add this patent to your site
Drop this plain-English patent card into any blog post or article — free, no signup. It always links back to the full breakdown here.
<div data-patentlens-widget data-patent-number="US12307349"></div> <script src="https://patentbrief.org/embed.js" async></script>
Stay in the loop
Get a weekly digest of new patents.
One email per week. No spam. Unsubscribe anytime.
Keep exploring
Related patents you should know
US 4683195 · 1987
How to Make Billions of Copies of a DNA Segment
This patent describes the Polymerase Chain Reaction (PCR), a method to rapidly create many copies of a specific piece of DNA or RNA, enabling its detection and analysis.
Cetus Corp
US 8697359 · 2014
How to Edit Genes in Human Cells Using an Engineered CRISPR System
This patent describes an engineered CRISPR-Cas9 system for precisely cutting DNA in eukaryotic cells to change how genes work, opening the door for gene editing in complex organisms.
Massachusetts Institute of Technology
US 7657849 · 2010
How the iPhone's Slide-to-Unlock Gesture Works
Apple's 2010 patent describes unlocking a device by dragging a specific graphical image across the touchscreen along a predefined path, a gesture that became iconic with the original iPhone.
Apple Inc
US 4733665 · 1988
How Doctors Implant a Permanent Stent Using a Balloon
This patent describes the method for placing a permanent, expandable wire mesh tube inside a blood vessel or other body tube using a balloon-tipped catheter to widen it and keep it open.
Expandable Grafts Partnership
US 4965188 · 1990
How to Make Many Copies of a DNA Piece with Heat
This patent describes the Polymerase Chain Reaction (PCR) method, a technique to make millions of copies of a specific DNA segment using a heat-resistant enzyme and repeated temperature changes.
Cetus Corp
US 4235871 · 1980
How to Encapsulate Active Materials in Lipid Bubbles Efficiently
This patent describes a method for trapping biologically active substances inside tiny, multi-layered fat bubbles called liposomes, using a specific water-in-oil emulsion and gel-forming process to improve how much material gets captured.
Individual
Semantically similar
You might also find these interesting
US 12353827 · 2025 · Unlikely Artificial Intelligence
Using Non-AI Systems to Improve AI Text Generation
US 12518214 · 2026 · Nant Holdings IP
Training AI on Private Data Without Seeing It
US 9548050 · 2017 · Apple Inc
How a Digital Assistant Launches Apps Using Your Voice
US 11204787 · 2021 · Apple Inc
How Digital Assistants Control Apps and Ask for More Information
More to explore
More in AI & Machine Learning
US 10452978 · 2019 · Google LLC
How AI Models Understand Language Using 'Attention'
US 6523026 · 2003 · Huntsman International LLC
How Computers Find Hidden Connections Between Different Fields of Knowledge
US 11615208 · 2023 · Capital One Services LLC
How Cloud Systems Automatically Create and Train AI Data Models
US 10402750 · 2019 · Facebook Inc
How Facebook Uses Deep Learning to Predict What You Might Like
New to patents?
Common Questions
Frequently Asked Questions
What does How an AI System Answers Financial Questions Using Specialized Bots cover?
This patent describes an AI system where a large language model (LLM) directs specialized machine learning agents to answer natural language questions about financial data, then refines the answers using an adversarial AI.
Who owns patent US 12307349?
Broadridge Financial Solutions owns this patent, granted in 2025.
When does this patent expire?
This patent is expected to expire on August 1, 2044, when the invention enters the public domain.
What problem does this patent solve?
This patent describes a method for AI systems to handle complex, domain-specific queries, particularly in finance, by combining the broad understanding of an LLM with the precision of specialized ML agents. This approach can improve the accuracy and reliability of AI-generated responses in critical fields. The inclusion of an adversarial agent for self-correction aims to address common issues of factual errors or 'hallucinations' in LLM outputs, making AI more trustworthy for professional use.
What does this patent NOT cover?
Does not cover systems that process natural language queries for non-financial data records, as Claim 1 specifies 'at least one attribute associated with at least one financial instrument'.
Patent monitoring