Background. Code review is critical for ensuring software quality in large-scaledevelopment, yet manual reviewers struggle to scale with growing codebases andorganizational complexity, particularly at large companies like Ericsson. Existingautomated review tools lack the depth needed for expert-level feedback.Objectives. This thesis introduces and evaluates an Assisted Code Reviewer (ACR)system, engineered to leverage project-specific context—such as design guidelines andcommit histories—to generate precise, actionable feedback. The goal is to enhancedeveloper productivity by reducing cognitive load, improve review efficiency acrosslarge teams, and bridge the divide between automated tools and human expertise inenterprise settings.Methods. Using Design Science Research, we first gathered information from rel-evant research and developers at Ericsson. We developed a solution, ACR, thatcombines generative ai-empowered agents along with knowledge graphs through aGraphRag approach. ACR was implemented and was statically evaluated on codechanges from two Ericsson products.Results. ACR effectively delivers valuable, context-enriched feedback, offering de-tailed line-specific comments that pinpoint issues and high-level summaries that con-textualize changes within project goals. Evaluations show that it adds practical util-ity, though its impact on code quality is inconsistent, occasionally missing the markcompared to senior reviewers.Conclusions. ACR provides a scalable, adaptable solution that enhances codereview by integrating seamlessly with version control systems, alleviating reviewerworkload while complementing human expertise.