HCLTech's Deployment Wall reveals 95% of AI pilots fail

HCLTech's Fabricio Costa has published a provocative paper arguing that enterprise AI's $37B investment boom is largely failing because organizations are focusing on the wrong problem. The research, published on arXiv as 'The Deployment Wall: A Diagnostic Framework and Instrument for Enterprise AI in the Deployment Era,' finds that 95% of enterprise generative AI pilots deliver no measurable profit-and-loss impact despite massive investments.

The paper introduces three key constructs: the Deployment Wall (a six-stage value-leak model), the Seam Index (a 0-12 diagnostic scoring platform readiness), and Deployment Debt (a quantifiable liability for unresolved friction). The framework shifts enterprise AI strategy from benchmarking models to evaluating architectural readiness, converting platform decisions from technical comparisons to architecture comparisons.

Costa argues that enterprise AI has entered a 'Deployment Era' where advantage comes not from model intelligence but from removing organizational and architectural friction that prevents capable models from reaching production. The research includes six falsifiable propositions and a scoring protocol with evidence anchors for consistent application across enterprises.