Causal Predictive Optimization and Generation for Business AI
Liyang Zhao, Olurotimi Seton, Himadeep Reddy Reddivari, Suvendu Jena, Shadow Zhao, Rachit Kumar, Changshuai Wei
TL;DR
CPOG addresses the challenge of optimizing B2B/SaaS sales by unifying causal prediction, multi-objective constrained optimization, and explainable Generative AI serving into a single end-to-end framework. The Prediction Layer uses causal uplift models and engagement forecasts to forecast the impact of sales actions; the Optimization Layer blends constrained mixed-integer programming with a contextual bandit to allocate efforts across accounts and reps while balancing revenue and engagement goals. The Serving Layer delivers transparent recommendations with template-based and GAI-generated explanations, supported by a feedback loop and human-in-the-loop design, and is demonstrated via LinkedIn deployments with strong lift and high user satisfaction. Offline analyses, ablations, and an observational online study validate the individual components and the integrated system, offering actionable guidelines for practitioners to implement causal predictive optimization in enterprise sales contexts.
Abstract
The sales process involves sales functions converting leads or opportunities to customers and selling more products to existing customers. The optimization of the sales process thus is key to success of any B2B business. In this work, we introduce a principled approach to sales optimization and business AI, namely the Causal Predictive Optimization and Generation, which includes three layers: 1) prediction layer with causal ML 2) optimization layer with constraint optimization and contextual bandit 3) serving layer with Generative AI and feedback-loop for system enhancement. We detail the implementation and deployment of the system in LinkedIn, showcasing significant wins over legacy systems and sharing learning and insight broadly applicable to this field.