Cognitive Solutioning of Highly-Valued IT Service Contracts

Abstract

Different service providers compete to win high valued IT service contracts in a tender kind of process, by providing comprehensive solutions that would fulfil the requirements of their client. Client requirements include services such as account management, storage systems, databases, and migrating the client infrastructure to the cloud. Preparing a solution is a step-wise process where a client shares the Request for Proposals (RFP) which documents the details of services required, and then service providers prepare solutions to fulfil these RFPs. The latter, usually referred to as “solutioning”, can be a lengthy, time-consuming process. Therefore, solutioning automation could result in efficiency increases and cost reductions for the providers. In this paper, we propose an automated, cognitive, end-to-end solutioning methodology that is comprised of 3 steps. The first step involves a textual analytics approach for mining the RFP documents, and extracting the client requirements and constraints. Second, we formulate an optimization model that chooses the optimal set of offerings (that the provider can offer) and their attribute values that cover the client requirements at a minimum cost. Finally, we require market benchmarks to compute pricing of the chosen offerings. Market benchmarks are sometimes unknown for some offerings or for some attributes of these offerings. Thus, in that third step, we show an iterative method to estimate the missing benchmarks along with a confidence score. We validate our methodology by applying it to a real-world application with a comprehensive dataset. We show that it takes minutes to run our method, compared to the days and sometimes weeks for the case of manual solutioning. We also illustrate that our methodology provides more accurate solutions compared to manual solutioning.