After specifying the grade level and class period, the AI still needs to know which references to use for this lesson. Context refers to the information the model can reference for this particular session; uploading materials, successfully reading them, and selecting them for the current task are separate things.

First, Select the Parts Relevant to the Task

For this Silk Road lesson, the student background, materials M01–M03, and learning objectives are needed. The school sports meet notice is unrelated to this lesson and should not be included. More materials do not equal more effective information; clearly mark which sources are evidence and which are merely style references.

MaterialHow to Use
Student backgroundControl knowledge difficulty and activity formats
M01, M02Support explanations of routes and exchanges
M03Organize discussion questions
An old lesson planReference the structure; do not copy its era or class directly into this lesson

After Providing, Confirm the Actual Reading Results

You can ask the model to summarize the purpose of each part of the materials and point out what is missing. More reliable evidence includes actual reading results, corresponding excerpts, or material numbers. A single sentence like "received attachment" or "already read" cannot prove that the entire materials are within the current reference scope.

The following three sections are the teaching basis for this session. First list what M01–M03 each support, then write the lesson plan. Any textbook page numbers not provided should be marked as pending verification.

When Materials Are Too Long, First Clarify the Focus

State the task, relevant sections, and questions to extract first. Applications may filter or compress materials, and models have a limit on how much they can reference in a single turn. List key facts separately and return to the original text to verify in the results.

If the application cannot read text from images, the attachment format is unsupported, or content is missing, resolve these issues first. Adding a longer prompt cannot substitute for successful reading.

When to Connect a Service

Using three materials at once may be sufficient with copy-paste or upload. When materials are continuously updated or frequently queried by number, supported applications can connect to a data service, such as using the Model Context Protocol (MCP). It standardizes communication, but whether materials are actually readable still depends on the actual returned results.

Next Steps

Continue learning how to respond to first-round feedback; for complete materials, read the Unified Lesson Planning Case Study.

Related Concepts

  • Context: Information the model can reference when processing the current task, such as requirements, conversation history, material excerpts, and tool return results.
  • Prompt: Your task request to the AI, which can be a plain English sentence or include background information, constraints, and output format.

General definitions and technical sources are on the related concepts pages. Verification date: 2026-09-09. The case study in this section is for instructional writing and optional for comprehension; it does not restrict continued reading.