United StatesEducation2024 – ongoingNorth Dakota University System
Policy Assistant for searching and answering questions about system policies
North Dakota University System sped up policy search 10 to 20 times with a generative AI assistant
Problem
The North Dakota University System spans 11 schools and serves about 80,000 students, faculty and staff. Its policies sit in a large set of documents, which made it slow to find and answer policy questions.
Approach
The system built Policy Assistant, a generative AI tool that searches policy documents and answers questions about them. It tested open-source language models, including Llama 2, with plans to move to DBRX. Over 3,000 public PDFs were loaded into the Databricks platform, and the tool was built in six months.
Results
Takeaway
A document-search assistant built on an organization's own policy files can cut lookup time sharply. The system did not say how it measured the speed-up.
Categories
- Size
- Large enterprise
- Industry
- Education
- Business function
- Internal work and knowledge management
- Innovation type
- Process innovation
- AI technology
- Generative AI
- Adoption
- Off-the-shelf tool or platform
- Evidence
- Vendor case study
Sources
Source summaries
Databricks describes how the North Dakota University System built Policy Assistant, a generative AI tool for searching and answering questions about policies. The page reports that using a large language model made policy search and response 10 to 20 times faster. It also says the tool was built in six months and draws on more than 3,000 public PDFs.
Full details are in the original sources.