At Anthropic, we believe that empowering people with AI, and ensuring that AI makes safe contributions to society, requires engaging with a wide range of human perspectives and experiences. Builders — product people, engineers, and makers — are uniquely positioned to leverage AI to amplify what they create—but only if they can approach it intentionally, with clear judgment and purpose.

This course helps builders develop practical AI collaboration skills through the 4D Framework (Delegation, Description, Discernment, and Diligence). In this course, we'll explore how the framework applies to real builder tasks—from framing a customer problem to shipping a product—and consider what it means to integrate AI responsibly across the full arc of building.

This course is built in partnership with CodePath, drawing on research with engineers and product builders about their actual needs, concerns, and aspirations for AI adoption.

The work builds on our AI Fluency: Framework & Foundations course, adapted specifically for the builder context where speed of implementation and the human role in judgment create unique considerations for AI collaboration.

Recommended prerequisites

This course lightly covers the foundational AI Fluency concepts. However, for deeper understanding, participants should complete AI Fluency: Framework & Foundations before beginning this builder-focused curriculum.

It is also recommended that learners have access to an AI tool for hands-on practice. Examples in this course will use Claude.ai, but any AI assistant will work.

Syllabus

Introduction and AI Fluency framework

Define AI Fluency and meet the four interconnected competencies, Delegation, Description, Discernment, and Diligence, that ground every productive AI collaboration in a builder's work. You'll build a reusable brief that hands AI your values, constraints, and context, learn to tell the inner loop from the outer loop, and map your current work to the 4Ds to find your biggest opportunity.

How AI works

Build a working mental model of what generative AI can and can't do reliably. You'll see how large language models like Claude are built, the developments that made them possible, and where they predictably fall short, so you understand why those limits matter for what you choose to build with AI.

The 4D framework for builders

Put Delegation and Description to work on real builder tasks: decide what to hand to AI and what judgment to keep yours, write acceptance tests that define done before a single line of code exists, and follow the Description Chain from a messy user need to a precise instruction AI can act on.

Discernment for builders

Apply Discernment to everything AI produces. You'll evaluate AI-generated code across five lenses and critique AI-built user experiences with feedback specific enough to act on, so "it works" stops being the bar and your taste becomes the differentiator.

Tying it all together

Bring Diligence full circle by owning what you ship, from whether it should exist to whether it's serving users after launch, and learn to make the call to ship it, fix it, or stop it with feedback loops and tests as your safety net. Then bring all four dimensions of the 4D Framework together, gather resources to keep building your AI fluency after the course, share what you've learned with colleagues, and complete the final course quiz to earn your certificate of completion.