Philosophy

We build agentic tools to amplify design research, not to automate it away. In the Design Research Collective, progress comes from tight loops between messy reality and clear evidence: framing problems with people, generating and testing ideas quickly, making assumptions explicit, and iterating with humility. Our agents exist to strengthen those loops.

  1. Human-first, always. The best research questions, judgments, and ethical calls stay with humans. Agents should reduce cognitive load, widen option space, surface blind spots, and help teams communicate, without flattening nuance or overriding expertise. We treat agency as delegation with accountability, not autonomy for its own sake.

  2. Evidence over vibes. Agents must be able to show their work: sources, provenance, uncertainty, and the chain of decisions that led to an output. If something can’t be traced, inspected, and debated, it doesn’t belong in a research workflow. Outputs should be reproducible, auditable, and easy to critique.

  3. Design is a process, not a snapshot. We value systems that support end-to-end research practice: literature sensemaking, study planning, data handling, analysis, interpretation, writing, and dissemination. Our agents should operate as scaffolding across the lifecycle, helping us move faster and think better, while preserving the context that gives results meaning.

  4. Tooling that respects reality. Research happens under constraints: time, budgets, compute, and human attention. Agents should be right-sized to the task, transparent about limits, and robust under imperfect inputs. They must integrate into existing tools and habits, not demand a total workflow rewrite.

  5. Collaboration is the unit of impact. Our work is inherently sociotechnical. Agents should make teams better: clearer handoffs, shared mental models, consistent documentation, and healthier critique. They should help us teach, mentor, and learn, supporting students as they become independent researchers, not substituting for that growth.

  6. Safety, privacy, and integrity are non-negotiable. We handle sensitive data, human subjects contexts, and high-stakes interpretations. Agents must minimize data exposure, enforce appropriate boundaries, and default to conservative behavior when uncertain. We would rather be slower than sloppy.

In short: we build agents that make our research more creative, more rigorous, more reproducible, and more human.