You have the ability to generate multiple chatbots tailored to various library activities, events, instructions, and more.
AgentBuilder
AgentBuilder is a web application demo empowered by AI. Its purpose is to enable public libraries to effortlessly craft and personalize AI chatbots, all without requiring any coding skills. It's a research demo created by researchers from UIUC iSchool.
AgentBuilder
AgentBuilder is a web application demo empowered by AI. Its purpose is to enable public libraries to effortlessly craft and personalize AI chatbots, all without requiring any coding skills. It's a research demo created by researchers from UIUC iSchool.
You have the ability to generate multiple chatbots tailored to various library activities, events, instructions, and more.
1. Prepare the information you what it to know.
Setting it up involves inputting certain information for its initial responses. You can use a PDF of 2-10 pages related to a patron-related context in the public library in this demo tryout. For example, a pdf about the library’s service, or events; a instruction on how to use a specific machine in the library.
Don’t forget to prepare some PDFs document during the user study try-out!
2. Information-Driven Agent Responses
The agent responds based on the information you have provided.
3. Train your agent
Should the initial response not meet your expectations, we provide a capability to enhance its intelligence through training and learning.
4. Test it Automatically
Interested in comprehending how your crafted chatbot responds across various contexts? Put it to the test with automated evaluations!
Q/A
Answer who / what / where / when questions
Answer questions about your library and its services
Instruct people on How-Tos
…
Diverse
Simulate chat for auto-testing
Translate to or from many languages
Support persona creations
…
About Us
We are a team of researchers from the University of Illinois at Urbana-Champaign, School of Information Sciences. This project is funded by Institute of Museum and Library Services (RE-251246-OLS-2022-IMLS). Questions? Contact research assistant: Qingxiao Zheng, or supervisor: Dr. Yun Huang.