Human-Centered Design (HCD) emphasizes user-centric research but faces inefficiencies and subjectivity in traditional methods. While AI enhances data processing in early and late research stages, existing tools remain passive, lacking real-time collaboration during interviews. This study explores AI’s active role in HCD through three questions: integrating AI into workflows, enhancing creativity, and defining human-AI collaboration models. Semi-structured interviews with 12 researchers across industries highlight gaps in real-time interaction, insight depth, and multimodal data integration. Key design requirements include improving communication, data quality, analytical efficiency, and research translation. Participants expect AI agents to automate processes, support decisions, and adapt to domains. We propose Co-Researcher, an AI framework enabling real-time transcription, dynamic questioning, and contextual analysis. The system fosters co-creative interviews, with AI suggesting follow-ups and generating structured reports. Results show Co-Researcher reduces manual tasks and deepens insights, positioning AI as a collaborative partner rather than a tool. This work advances human-AI collaboration in HCD, balancing automation with creativity and rigor, while addressing cross-functional alignment and methodological challenges.