How Amazon is using AI for Enterprise Research at Scale

AI has moved through exploration and adoption — and enterprise research teams are now facing the harder question: how do you actually scale it? In this session, Amazon's Sr. UX Researcher Sarah Kling and Knit CEO Aneesh Dhawan trace Amazon's three-year journey from cautious AI experimentation to a model where human expertise and AI work in true partnership. They'll share what went wrong early on (hallucinations, weak synthesis, outputs you couldn't trust), what changed when they committed to a hybrid human-plus-AI approach, and what it looks like when research stops being a deliverable and starts shaping decisions before they're made. Through three real examples — including a five-day turnaround that moved a strategic initiative forward at the senior leadership level — attendees will see what enterprise-scale AI research looks like in practice, not just in theory.

Takeaways:

• Why AI speed without research rigor creates faster ways to be wrong — and what the human-plus-AI model actually fixes
• How combining quant and qual through a hybrid approach produces richer, more trusted insights than either method alone
• What it means for research to "move upstream" — from reporting on decisions to shaping them in real time