department-of-veterans-affairs / va.gov-cms

Editor-centered management for Veteran-centered content.
https://prod.cms.va.gov
GNU General Public License v2.0
99 stars 69 forks source link

Content Audit: Chatbot NLU Model Training Phrases #17136

Open ksk385 opened 9 months ago

ksk385 commented 9 months ago

Content audits can be performed by searching the codebase and compiled site for specific terms ur using an audit tool within drupal. Please provide any context for the audit, any related terms or alternate spellings, or and related components that could return content you are interested in. Please see the collab cycle doccumentation for more information.

1. Please provide any context around your audit

As we look into ways to improve the intent recognition of the VA Chatbot, we would like to get a data set of all the questions on the VA.gov site along with their page titles and URLs.

2. What are the terms you are searching for within the CMS? Are there alternate spellings?

Everything thats an FAQ. Usually they are h2s but also accordion style questions as you scroll down on many pages.

3. Are there components, elements or content types related to your search?

Interested in the content rather than elements.

4. Does a corresponding audit need to be run on the Veteran-facing website?

5. When do you need the audit to be completed

No deadline as such but we are actively working on NLU improvements and this data would give us more to work with. If there were APIs we could call to get this data ourselves that would be ideal!

Team

Please check the team(s) that will do this work.

jilladams commented 9 months ago

Thread with context: https://dsva.slack.com/archives/C52CL1PKQ/p1706899196119909

This issuetype defaults to CMS team (fyi @EWashb ). The thread above pertains to a Public Websites/ CAIA project, so depending on Dave's thoughts, y'all could opt to ask PW for an assist here.