Rethink business value for generative AI use cases
summary
Claims and numbers: $22M savings for NIB with 60% chat deflection (voice ~15%), Orion Health folds 300,000 documents into an intelligent search for ~20% productivity gains, Deputy reports ~70% developer efficiency improvement. Amazon Q Business (launched Nov 2023, Sydney) connects 40+ sources; Bedrock and SageMaker are used to choose and fine‑tune models. Compute costs drop ~10x per year (cost halves every 4–6 months), pilots are common (14% skip ROI, ~21% run pilots), and levers like fine‑tuning (≈75% cost cut), prompt routing (≈30%), prompt caching (≈90%) and batching (≈50%) materially lower inference cost.
AI-written summary · about
chapters (7)
- 0:00Introduction and setup
- 2:56Reframing ROI for generative AI
- 5:47Tangible value from AI use cases
- 12:34Intangible benefits and risk reduction
- 18:56Investment strategies and cost allocation
- 26:00Optimizing AI model inference costs
- 30:17Key takeaways and closing
related talks
23mGenerative AI use cases
25mMaking AI Useful
33mShow Me The Money: Ensure AI Delivers Business Value
27mSoftware engineering with GenAI
20mGetting business value from AI: Everyday AI, extraordinary people
36mSolve common business problems with AI/ML
24mState of Startups and AI 2025
29mSolve common business problems with AI and ML