Developing a Base Framework to Enable a Lean Step Zero toward Dispute Avoidance

Abstract Lean construction techniques are being discussed for their benefits in dispute avoidance, along with waste reduction, productivity improvement, and minimizing time and cost overruns. However, implementing lean presents challenges, with stakeholder resistance to change being a primary obstacle. To address this resistance, current literature recommends adding new contract clauses to the traditional standard form contracts (SFCs) to make lean adoption contractually binding. However, since traditional SFC in construction often contains restrictive and one-sided clauses, integrating lean-friendly, collaborative clauses may conflict with existing ones, potentially creating new conflicts. This study, therefore, aims to assess the adaptability of Indian public sector undertakings (PSUs) SFCs to lean, i.e., the leanness of the contract. Built on a base framework developed from lean construction and contract drafting principles recommended in the literature, an automatic leanness assessment application is developed using a multilayer perceptron (MLP) model augmented with synthetic data augmentation and the synthetic minority oversampling technique (SMOTE), achieving an F1 score of 93%. After the assessment, the restrictive clause segments are neutralized using a lean-neutral template via a retrieval-augmented generation (RAG)- assisted large language model (LLM). This makes the traditional contract conducive to adding clauses that actively support lean implementation without conflicting with existing clauses.