Automated RFP Response Clustering for Search Accuracy

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing systems for responding to Requests for Proposal (RFPs) are inefficient due to manual content tagging processes, leading to inaccurate and time-consuming search outcomes across multiple content management systems and databases.

Innovation Solution

A computer-implemented method and system that uses search algorithms and clustering techniques to automatically determine relevant responses to RFP queries by grouping similar historical responses and augmenting them with content management system records, sorting results by relevance, and recommending the most appropriate answer.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If manual content tagging process is used to classify and prioritize content, then content can be organized in content management systems, but the process is inefficient and leads to mixed search outcomes

Engineering Contradiction:

Improvecontent classification efficiencyVSAvoidsearch outcome accuracy

Core Design Contradiction:

Ease of operationVSReliability

Solution Approach 1:

The patent replaces the manual mechanical tagging process with an automated machine learning system that uses natural language processing and clustering algorithms to classify and prioritize content, eliminating human labor while improving consistency and accuracy of search outcomes

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system enables content to automatically tag and classify itself through machine learning models that analyze content characteristics, allowing the content management system to self-organize without human intervention and provide reliable search results

Inventive Principle:

Principle #25Self-service

2Quantity of substance

If thousands of historical responses are stored in data repositories, then comprehensive answer options are available, but users are overwhelmed with excessive options

Engineering Contradiction:

Improveavailability of response optionsVSAvoiduser decision complexity

Core Design Contradiction:

Quantity of substanceVSEase of operation

Solution Approach 1:

The system extracts only the most relevant responses from the large repository of historical data by using clustering algorithms to identify and retrieve top-k similar responses, presenting a manageable subset to users while maintaining comprehensive coverage of available options

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The patent segments the large set of historical responses into clustered groups based on similarity, organizing thousands of options into manageable categories that users can easily navigate and select from, reducing cognitive load while preserving answer diversity

Inventive Principle:

Principle #1Segmentation

3Quantity of substance

If multiple content management systems and databases are used to store RFP content, then comprehensive content storage is achieved, but searching across systems becomes time-consuming

Engineering Contradiction:

Improvecontent storage capacityVSAvoidsearch time

Core Design Contradiction:

Quantity of substanceVSLoss of time

Solution Approach 1:

The patent merges multiple content management systems and databases into a unified search interface that accepts single queries and automatically searches across all connected systems, retrieving and consolidating results from multiple sources simultaneously to eliminate manual system-by-system searching

Inventive Principle:

Principle #5Merging (Combining)

4Productivity

If automated tagging is used based on user search patterns, then tagging efficiency is improved, but accuracy decreases due to random user behavior patterns

Engineering Contradiction:

Improvetagging automation speedVSAvoidtagging accuracy

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The patent replaces automated tagging based on random user behavior patterns with machine learning models that use natural language processing and semantic analysis to determine accurate content tags and classifications, achieving both automation and precision

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentUS12086170B2Systems and methods for searching content to queries

Publication Date: 2024.09.10 FMR CORP

AI summary

A computer implemented method is provided for automatically determining a response to an input query. The method includes searching at a first stage a data repository using the input query. The data repository is configured to store historical queries and their corresponding responses. The search is adapted to determine a plurality of historical queries related to the input query and historical responses corresponding to the related historical queries. The method also includes clustering the plurality of historical responses into one or more response groups, where each response group includes one or more similar historical responses. The method additionally includes searching at a second stage a content management system (CMS) to determine if there is at least one similar response to at least one historical response in each response group. The CMS is configured to store standard responses to anticipated queries.