Deep Research Agents for Shell Company Screening

Shell company screening is a complex enterprise research task for identifying and evaluating listed companies as potential shell acquisition targets. However, the task is challenging because analysts must search a large and changing market, adapt quickly as deal requirements change, and repeat due diligence across multiple candidates. We therefore study how Deep Research Agents can support this task in practice. We formulate shell company screening as a multi-stage research process and design a Deep Research workflow to carry it out. We examine the workflow through a real-world case of screening Hong Kong-listed companies. In this case, the agent carries out the major stages of the workflow, including candidate search, due diligence, transaction analysis, and prioritization. The case also reveals limitations in public information coverage, multi-source reasoning, and claim–evidence traceability. These findings highlight the capabilities and limitations of current Deep Research Agents in complex enterprise research tasks.