LSLAsks

A chatbot prototype from Lone Star Legal Aid to provide staff with centralized, searchable access to internal policies, benefits information, and workplace procedures.
Project Description
LSLAsks is a proposed internal information tool for Lone Star Legal Aid, using the same Retrieval-Augmented Generation (RAG) architecture as the Juris chatbot but with a different document scope. The chatbot would draw on a curated knowledge base that includes benefits documents, the employee handbook, internal forms, IT procedures, and HR policies.
The aim is to improve staff access to accurate, current workplace information. Currently, employees often need to search across multiple drives and systems, where documents may be outdated or duplicated. This can be particularly challenging during onboarding or when handling administrative tasks.
Development to date has included establishing a designated webpage for the tool and conducting workflow mapping sessions across departments. These sessions have identified high-need information types and common search scenarios. Staff interviews are ongoing to prioritize content for integration.
The proposed chatbot would allow employees to enter questions in plain language—e.g., “What is the process for requesting parental leave?”—and receive direct answers with links to relevant source documents. This could reduce routine inquiries to HR and operations staff, help maintain consistency in policy interpretation, and provide 24/7 access to essential resources.
Functional Scope
- Knowledge Base: Benefits documentation, HR and IT policies, workplace guidelines, forms, internal procedures
- Architecture: Modular RAG framework, adaptable to different knowledge domains
- Access: Web-based interface for staff across all LSLA offices
- Content Management: Curated, regularly updated source documents to ensure accuracy
- Integration Potential: Future expansion to include other internal knowledge areas or external service information