Abstract
Cost estimation is a critical preconstruction function increasingly targeted for artificial intelligence (AI)-enabled improvement, yet empirical evidence on how professional estimators integrate these tools into daily practice remains limited. This exploratory qualitative study examines AI adoption in construction cost estimation through semi-structured interviews with 12 professional estimators from eight U.S.-based cost consultancies and general contractors, supplemented by a corroborating interview with a senior technology executive. Drawing on the Technology Acceptance Model (TAM) and Diffusion of Innovations (DOI) theory as complementary frameworks, the analysis operationalizes seven theoretical constructs at the coding level and identifies six practitioner-validated priority themes: reliability and uncertainty signaling, workflow integration, systems coverage for mechanical, electrical, and plumbing (MEP) assemblies, market-linked pricing, document and scope intelligence, and human-centered automation. The study introduces and formally defines the verification paradox, a previously unnamed mechanism in which estimators must re-perform manual takeoffs to validate AI outputs, neutralizing efficiency gains. The six themes are operationalized into the Construction AI Adoption Readiness Assessment (CAARA), an 18-item diagnostic instrument mapped to a three-phase implementation roadmap, presented as a conceptual framework pending psychometric validation. This study contributes practitioner-grounded empirical evidence and a structured diagnostic tool to support phased AI integration in professional estimation contexts.