Choose the Smallest Surface That Can Carry the Work
Product selection is architecture at knowledge-work scale. The wrong surface can make correct output stale, unreviewable, or needlessly expensive. Choose among chat, Projects, Research, files and Artifacts, connectors, and programmatic surfaces. Explain the durable role of Haiku, Sonnet, and Opus without depending on a specific model version. Match a surface and model to quality, speed, cost, freshness, and governance constraints. Compare direct Anthropic, Amazon Bedrock, Google Vertex AI, and Microsoft Foundry deployment paths with an architecture decision record. Identify when memory, project knowledge, or a new conversation is the correct continuity mechanism. Mark changeable product facts with an official source and verification date. An operations lead prepares a weekly competitor brief. She opens last week's chat, pastes three new links, asks for an update, and forwards the result. The output is polished. It also cites an old product price from the previous conversation, misses a policy change in an internal document, and contains a competitor claim with no source. The failure did not begin with wording. It began with the chosen work surface. An old chat carried stale context. Pasted links did not guarantee comprehensive research. The internal policy was not part of the available knowledge. The workflow had no claim-verification step. The exam objective is called product and model selection, but the real skill is boundary design. You decide…
Choose the Smallest Surface That Can Carry the Work: Product selection is architecture at knowledge-work scale. The wrong surface can make correct output…
This free lesson is part of the AI Engineering from Scratch curriculum. Read the full explanation, run the lesson code, and verify the result in the interactive reader or from the repository source.
Browse the complete course catalog or open this lesson on GitHub.