AI Prompting course - Lesson 2 of 10

Lesson 2: Google Gemini: Give multimodal tasks and structured instructions

Learn practical prompts for ChatGPT, Gemini, Claude, Microsoft Copilot, Perplexity, GitHub Copilot, Adobe Firefly, Canva, Grammarly, and NotebookLM.

← All AI Prompting lessons

Lesson 2 of 10 ยท Google Gemini

Give multimodal tasks and structured instructions

Where the version you use supports it, Gemini can work with text and other input types such as images or files. State exactly what you supplied and what you want inspected. For an image, ask for visible details rather than guesses about identity, intent, or context. For a search or factual answer, request sources and verify them independently.

Example: inspect a product photo

Describe the attached product photo for an accessible catalog entry.
Return: (1) a factual description of visible objects and colors,
(2) details that cannot be determined from the image.
Do not infer a brand, material, dimensions, or product claims unless
they are visibly provided. Maximum 70 words.

Privacy check: Remove faces, addresses, account details, or other sensitive information from an image before uploading unless you have permission and the service is approved for that data.