It doesn’t need to be a chatbot: Unlocking the product value of smaller AI models
summary
Argues that product AI succeeds when teams pick specific, small wins instead of defaulting to big chatbots, because integration gaps and noisy LLM capabilities create real risks and costs. Recommends starting with smaller language models (a few billion parameters), using internal enablers like analytics, and building features such as Booking.com’s smart filter translation to structured query. Presents an AI opportunity tree (insights, remove friction, add functionality, personalize) and cites ~80% failure rates to justify incremental rollout.
AI-written summary · about
chapters (5)
- 0:00Smaller AI models
- 1:05Reality check on AI
- 5:38Why start small
- 7:22AI opportunity tree
- 13:15Building small AI features
related talks
28mBuilding a Micro AI company in 2017
29mAvoiding the Chatbot Trap Designing AI Interfaces That Unlock Real Business Value
27mBuilding AI applications: Successes and Lessons
23mDelivering business success with AI
37mPanel Discussion: Developing with AI
25mConversational AI: Sales and Customer Service Chatbot Use Cases for Small and Medium Sized Business
36mThe Best AI Strategy? Don’t Build AI 🤯 Build AI-Ready Software
01h 04mBuilding Expertise Into Generated Conversations