Awkward Silences Research Ops 2.0, Episode 2: Leveraging AI for Research Superstardom
21 snips
Sep 2, 2025 Dennis Meng, Co-founder and CPO at User Interviews, and Lexi Brights, Director of Research at Workday, dive into the evolving world of Research Ops shaped by AI. Meng elaborates on how research operations can harness AI agents, while Brights shares insights on utilizing AI to derive meaningful data from open-text survey responses. They discuss the balance between human oversight and AI autonomy, the challenges researchers face, and how continuous learning is essential in this rapidly advancing field. Expect engaging case studies and thought-provoking debates on AI's future role in research.
AI Snips
Chapters
Books
Transcript
Episode notes
AI As A Third Constituent
- ResearchOps must treat AI as a new constituent that needs governance and controls.
- Dennis Meng warns organisations to plan how autonomous AI agents will be kept under control.
When LLMs Invent Quotes
- Lexi Brights fed an LLM open-text survey responses and asked for quotes, but the model invented charming quotes that didn't exist.
- She discovered the LLM 'just thought that would be more interesting' and learned to enforce data-presence constraints.
Be Precise With Prompts
- Learn prompt engineering because the questions and constraints you give an LLM determine its outputs.
- Lexi urges explicit instructions like 'only return quotes present in the dataset' to avoid hallucinations.
