6 - Business Operations Streamlining System (BOSS)

The objective of the Business Operations Streamlining System (BOSS) project is to leverage artificial intelligence (AI) and machine learning (ML) techniques to help Navy program managers (PMs) keep complex programs on schedule and on budget.  Specifically, BOSS will monitor and interpret internal and external documents to track program progress, using knowledge about schedules, budgets, technologies, and contractors to recognize potential problems early and help frame effective responses swiftly.  Automated data processing will winnow the overwhelming volume of information PMs must track within and across programs.  Early problem recognition will enable PMs to retain a wider range of lower-cost and -risk response options, such as parallelizing or expediting the most flexible tasks.  Support for generating and evaluating responses will promote effective low-impact options. We are addressing three central PM challenges: (1) too many problems revealed (or hidden) in too much data; (2) fewer options with worse outcomes available for late responses; and (3) consequences obscured across program management boundaries.  Five technologies meet the need: (a) Model-Based Knowledge Management, (b) domain-specific information extraction and analysis, (c) diagnostic and prognostic decision assistance, (d) case-based reasoning, and (e) knowledge-driven scheduling and resource allocation.  We have unique experience and assets for all of these, including the TEAM, Aurora, and BEACON systems.  Our Phase II team and plan will deliver a validated solution to real Navy problems, saving millions of dollars and keeping programs on track for success. The primary goal of Phase II will be to develop, test, and evaluate a complete BOSS operational prototype at Technology Readiness Level (TRL) 6—preparing for Phase III TRL 9 transition.  Phase II will employ iterative development and agile techniques.  The first iteration will focus on development infrastructure, system architecture, major services, and the domain model. Iteration two will introduce data monitoring and analysis workflows, while elaborating knowledge and heuristics for the scheduler.  The third release will include some version of essentially all major Phase II features, enhancing data trend analysis and text extraction, and introducing new problem/response analysis capabilities.  The fourth and final release will be a feature-complete prototype ready for final evaluation.  Phase II work will include a broad range of Subject Matter Experts (SMEs) and data: We will continue to work with our retired Navy SMEs; we will seek input from active Navy programs and PMs; and we will work with PMs at Northrop Grumman. Northrop Grumman will also provide historical data to drive system development and, later, test and evaluation; our close work with them will facilitate commercialization.