Premium membership
Continue this concept with Premium
Premium includes all published courses, graded answers, concepts and PDF guides. See the price and access terms on the pricing page.
View Premium accessone payment · six months of site access
Advanced
Deterministic Plus Probabilistic: Designing Workflows With Rules, Models and People
Reliable AI products combine three kinds of component: deterministic code for policy, calculations and money movement; models for messy input, language and pattern recognition; and people for judgment and exceptions. This guide shows how to decompose a workflow into those parts, with an expense-assistant example, the interfaces between them and the mistakes that blur the lines.
CONCEPT / AI PRODUCT DISCOVERY
RELATED CONCEPTS
PRACTISE THIS CONCEPT
LLM & RAG FundamentalsOur CEO asks whether we should fine-tune a model on our documents or use RAG. Explain the difference and make a recommendation.→Evals & Launch DecisionsYour support agent completes 82% of tasks in testing. Is it ready to launch?→Agentic ProductsYour team wants to build an AI agent to onboard new suppliers. Should it be an agent, or something simpler?→AI Product SenseA stakeholder wants to 'add AI' to our expense approval flow. How do you decide whether it should use AI at all?→AI UX & Human-in-the-LoopWhat do people mean by 'levels of autonomy' for AI features, and how would you choose the right level?→AI System DesignWhy would a company put a gateway between its applications and the model providers it uses?→