Beyond diagnostics: a scoping review of generative artificial intelligence for clinical documentation and client communication in veterinary practice

Abstract Objective To systematically map the current landscape of generative AI (GenAI) and large language model applications for clinical documentation and client communication in veterinary medicine, critically appraise validation depth, and identify barriers to real-world adoption. Methods Following Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for scoping reviews guidelines, PubMed, Scopus, CAB Abstracts, and IEEE Xplore were searched for original research published between January 2020 and June 2026. Two reviewers independently screened records and extracted data using a 4-dimensional validation matrix (technical accuracy, clinical quality and safety, operational impact, and human-centric impact). Synthesis was conceptually grounded in the job demands-resources model. Results 10 studies met the inclusion criteria: 6 addressed clinical documentation (automated clinical coding, information extraction, and generation of subjective, objective, assessment, and plan [ie, SOAP] notes), and 4 addressed client communication (veterinary triage and automated query responses). While technical accuracy was frequently assessed, operational impact was evaluated in only 1 study, and no study employed validated psychometric instruments to measure burnout. Key barriers included clinical hallucinations, data privacy concerns, poor interoperability with practice information management systems, and absence of veterinary-specific fine-tuning. Conclusions GenAI demonstrates substantial technical promise for automating veterinary documentation and client communication tasks, yet the evidence base is characterized by a pronounced validation gap: operational and human-centric outcomes remain virtually unquantified. Clinical Relevance Until rigorous evidence emerges, veterinarians should approach commercially available AI tools with cautious optimism. Future research must prioritize pragmatic, multicenter prospective trials employing multidimensional validation frameworks to generate evidence supporting safe clinical deployment.