Year: 2025
Type: article
Abstract: Enterprise Question-Answering (Q&A) systems have emerged as critical components in modern business infrastructure, revolutionizing how organizations manage and utilize their information assets. These systems integrate sophisticated retrieval mechanisms with advanced language models while maintaining robust privacy frameworks. The implementation encompasses multiple architectural strategies, from large-scale model training to hybrid retrieval-augmented generation frameworks, each offering distinct advantages for enterprise deployments. Through the integration of domain-specific understanding pipelines and dual-index retrieval strategies, these systems achieve superior query comprehension and response accuracy. The Plan-Act framework facilitates effective coordination between retrieval mechanisms and language models, while comprehensive security measures ensure data protection through multi-layered access controls and isolation strategies. The incorporation of automated governance tools and continuous compliance monitoring establishes a foundation for secure, scalable enterprise Q&A deployments. These systems demonstrate remarkable effectiveness in handling diverse business domains, from IT support to HR queries, while maintaining stringent privacy standards and regulatory compliance.
Author: Nitish Ratan Appanasamy -
Institution: Columbia University
Language: English
Topic: Topic Modeling
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W4409218695