A RFP System for Generating Response to a Request for Proposal
Asha Rajbhoj, Padmalata V. Nistala, Vinay Kulkarni, Gomathi Ganesan · 2019
Responding to Request for Proposal (RFP) with comprehensive solutions is central to IT Services business. Typically, an RFP comprises a set of questions spanning across various domain areas. Current industry practices largely rely on Subject Matter Experts to analyze questions and search multiple information sources to come up with the best possible response for a question keeping in view the customer context. With expertise typically in short supply, it becomes increasingly difficult to manage growing RFP volumes. To address this problem we propose a generic solution meta-model and a system that can generate a draft response to an RFP to be augmented manually later. The generation step uses NLP, modelling and search techniques augmented with a knowledge base to generate knowledge search queries to retrieve suitable answers to a set of RFP questions and compose RFP response. In this paper, we share a RFP response generation approach, its implementation, results and lessons learnt from deployment with one of the business units having high volumes of proposal turnover. Initial results show that RFP system is effective in enabling response generation with ~76% mean query precision and 86% mean query recall.