# US 2026/0089158 A1 · Web accessibility monitoring and correction using ai

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

- Number: US 2026/0089158 A1
- Document ID: US20260089158A1
- Status: Published application
- Kind: A1
- Application number: 18/896,869
- Filed: 2024-09-25
- Published: 2026-03-26
- Earliest claimed: 2024-09-25
- Assignees: International Business Machines Corporation
- Applicants: International Business Machines Corporation
- Inventors: Willie L. Scott, II; Michael Brooke Gower; Mary Jo Mueller; Phillip D. Jenkins; Marc Kevin Johlic; Marnie Hoover
- Web page: https://patentgrep.com/patent/US20260089158A1

## CPC

- H04L 63/10: {for controlling access to devices or network resources}
- H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M)

## Abstract

A computer hardware system includes one or more artificial intelligence (AI) foundation models. Using a collection agent, web accessibility incident data associated with web assets is scraped from a plurality of different sources. The web accessibility incident data is classified, by an AI classifier, into a plurality of differently-classified groups of data. Entities are extracted by an AI entity extractor from the plurality of differently-classified groups of data. Using a frequency analysis engine and from the extracted entities, a frequency distribution table is created that identifies how frequently types of non-compliance are contained within the web accessibility incident data. Using the frequency distribution table a web accessibility meta-analysis is generated by an ensemble engine, and the web accessibility meta-analysis is provided to an accessibility checker, which modifies the operations thereof. The one or more AI foundation models includes the AI classifier and the AI entity extractor.

## Claims

### Claim 1

A computer-implemented method including one or more artificial intelligence (AI) foundation models, comprising:
scraping, from a plurality of different sources and using a collection agent, web accessibility incident data associated with web assets;
classifying, by an AI classifier, the web accessibility incident data into a plurality of differently-classified groups of data;
extracting, by an AI entity extractor, entities from the plurality of differently-classified groups of data;
creating, using a frequency analysis engine and from the extracted entities, a frequency distribution table identifying how frequently types of non-compliance are contained within the web accessibility incident data;
generating, using the frequency distribution table and by an ensemble engine, a web accessibility meta-analysis; and
providing the web accessibility meta-analysis to an accessibility checker, wherein the web accessibility meta-analysis is used to modify operations of the accessibility checker, and
the one or more AI foundation models includes the AI classifier and the AI entity extractor.

### Claim 2 (depends on 1)

The method of claim 1, wherein
the web accessibility incident data is classified by type of compliance violation.

### Claim 3 (depends on 1)

The method of claim 1, wherein
the extracted entities include web accessibility standards contained within the web accessibility incident data.

### Claim 4 (depends on 1)

The method of claim 1, further comprising
generating, using a AI summarizer, a summary for each of the plurality of differently-classified groups of data, wherein
the web accessibility meta-analysis is generated using the summary for each of the differently-classified groups of data, and
the one or more AI foundation models includes the AI summarizer.

### Claim 5 (depends on 4)

The method of claim 4, wherein
the entities extracted by the AI entity extractor are used by a web accessibility standards agent to identify documentation describing web accessibility standards associated with the entities, and
the AI summarizer uses the documentation to generate the summary.

### Claim 6 (depends on 1)

The method of claim 1, wherein
a web asset analyzed by the accessibility checker is subsequently modified based upon an analysis performed by the accessibility checker.

### Claim 7 (depends on 1)

The method of claim 1, wherein
a web asset is a publicly-accessible web page or portion thereof.

### Claim 8

A computer hardware system including one or more artificial intelligence (AI) foundation models, comprising:
a hardware processor configured to initiate the following executable operations:
scraping, from a plurality of different sources and using a collection agent, web accessibility incident data associated with web assets;
classifying, by an AI classifier, the web accessibility incident data into a plurality of differently-classified groups of data;
extracting, by an AI entity extractor, entities from the plurality of differently-classified groups of data;
creating, using a frequency analysis engine and from the extracted entities, a frequency distribution table identifying how frequently types of non-compliance are contained within the web accessibility incident data;
generating, using the frequency distribution table and by an ensemble engine, a web accessibility meta-analysis; and
providing the web accessibility meta-analysis to an accessibility checker, wherein
the web accessibility meta-analysis is used to modify operations of the accessibility checker, and
the one or more AI foundation models includes the AI classifier and the AI entity extractor.

### Claim 9 (depends on 8)

The system of claim 8, wherein
the web accessibility incident data is classified by type of compliance violation.

### Claim 10 (depends on 8)

The system of claim 8, wherein
the extracted entities include web accessibility standards contained within the web accessibility incident data.

### Claim 11 (depends on 8)

The system of claim 8, wherein the hardware processor is further configured to initiate the following executable operation:
generating, using a AI summarizer, a summary for each of the plurality of differently-classified groups of data, wherein
the web accessibility meta-analysis is generated using the summary for each of the differently-classified groups of data, and
the one or more AI foundation models includes the AI summarizer.

### Claim 12 (depends on 11)

The system of claim 11, wherein
the entities extracted by the AI entity extractor are used by a web accessibility standards agent to identify documentation describing web accessibility standards associated with the entities, and
the AI summarizer uses the documentation to generate the summary.

### Claim 13 (depends on 8)

The system of claim 8, wherein
a web asset analyzed by the accessibility checker is subsequently modified based upon an analysis performed by the accessibility checker.

### Claim 14 (depends on 8)

The system of claim 8, wherein a web asset is a publicly-accessible web page or portion thereof.

### Claim 15

A computer program product, comprising:
a computer readable storage medium having stored therein program code,
the program code, which when executed by a computer hardware system including one or more artificial intelligence (AI) foundation models, causes the computer hardware system to perform:
scraping, from a plurality of different sources and using a collection agent, web accessibility incident data associated with web assets;
classifying, by an AI classifier, the web accessibility incident data into a plurality of differently-classified groups of data;
extracting, by an AI entity extractor, entities from the plurality of differently-classified groups of data;
creating, using a frequency analysis engine and from the extracted entities, a frequency distribution table identifying how frequently types of non-compliance are contained within the web accessibility incident data;
generating, using the frequency distribution table and by an ensemble engine, a web accessibility meta-analysis; and
providing the web accessibility meta-analysis to an accessibility checker, wherein
the web accessibility meta-analysis is used to modify operations of the accessibility checker, and
the one or more AI foundation models includes the AI classifier and the AI entity extractor.

### Claim 16 (depends on 15)

The computer program product of claim 15, wherein
the web accessibility incident data is classified by type of compliance violation.

### Claim 17 (depends on 15)

The computer program product of claim 15, wherein
the extracted entities include web accessibility standards contained within the web accessibility incident data.

### Claim 18 (depends on 15)

The computer program product of claim 15, wherein the computer hardware system is further configured to perform:
generating, using a AI summarizer, a summary for each of the differently-classified groups of data, wherein
the web accessibility meta-analysis is generated using the summary for each of the differently-classified groups of data, and
the one or more AI foundation models includes the AI summarizer.

### Claim 19 (depends on 18)

The computer program product of claim 18, wherein
the entities extracted by the AI entity extractor are used by a web accessibility standards agent to identify documentation describing web accessibility standards associated with the entities, and
the AI summarizer uses the documentation to generate the summary.

### Claim 20 (depends on 15)

The computer program product of claim 15, wherein
a web asset analyzed by the accessibility checker is subsequently modified based upon an analysis performed by the accessibility checker.
