Mixture of Experts Copilot usage reveals AI adoption patterns
13 snips
Feb 13, 2026 Volkmar Uhlig, CTO and VP of Data Platforms, offers engineering and product takes. Lauren McHugh Olende, AI Open Innovation lead, brings enterprise prompting and adoption insights. Kush Varshney, IBM Fellow, covers governance and societal angles. They discuss Copilot usage patterns, differences between search and copilots, the Ralph Wiggum prompting tactic for coding agents, cost and licensing tradeoffs, and AI’s mainstream signals like Super Bowl ads.
AI Snips
Chapters
Transcript
Episode notes
Copilot Usage Mirrors Search With Personal Twist
- Microsoft Copilot usage mirrors search patterns but shows stronger sensitivity and seasonality.
- Users ask more personal, precise questions like health advice and holiday planning.
Enterprise Licenses Blur Work-Personal Lines
- Enterprise copilot use blurs work and personal queries, raising cost and privacy questions.
- Organizations will need policies or billing controls to manage who pays for non-work usage.
Try Minimal Looping Prompts For Agents
- Use simple repetitive prompts (Ralph Wiggum) to avoid constant babysitting of coding agents.
- Front-load clear instructions to reduce interactive interruptions and wasted loops.