What is an AI Prompt?

TL;DR

• An AI prompt is the input that frames what the model should do.
• A useful prompt can include a task, context, constraints, examples, and output format.
• Prompt quality matters, but context quality often matters more for complex product work.
• Longer prompts are not automatically better.
• Iterative prompting can refine a result, but persistent product context reduces repeated explanation.

What does a prompt do?

A prompt can be as short as “Summarize this interview” or as structured as a detailed product-design brief containing user goals, existing flow, constraints, design-system rules, edge cases, and expected output.

The prompt shapes what the model pays attention to during the current task.

What can a strong prompt include?

• Task: what should be done?
• Context: what should the model know?
• Constraints: what must or must not happen?
• Examples: what does good output resemble?
• Output format: how should the result be structured?

Prompt vs. context engineering

A prompt is one input. Context engineering is the broader practice of assembling the right information, tools, memory, and constraints for the model at the moment of execution.

Complex workflows usually cannot fit their entire operating context into one handcrafted prompt.

Why do prompts fail?

• the task is ambiguous
• important context is missing
• constraints conflict
• the request is too broad
• the expected output is undefined
• the model lacks access to the relevant source of truth

Common mistakes

Trying to fix missing context with more adjectives

“Make it premium and intuitive” cannot replace actual product rules.

Overconstraining creative exploration

Specify what matters and leave room where exploration is useful.

Assuming the prompt persists automatically

Different systems handle memory and context differently.

The bottom line

An AI prompt asks: what should this model understand and do in this turn, and what boundaries make the output useful?

Related terms

Generative AI · AI agent · Context engineering · Product context

Relevant Figr resource

Read AI for UX Design for examples of prompting inside context-aware product workflows.