Few-Shot Prompting: Teach by Example
Get consistent output by showing the AI examples instead of describing what you want.
Assigning the right role changes everything. Here's how to do it well.
ChatGPT Research Prompts for Deep Work
ChatGPT will hand you a confident paragraph and a made-up statistic with the same tone. Here's why "research this for me" fails, how to make it show its confidence instead of hiding it, and annotated prompts for scoping, sourcing, and fact-checking real research.
ChatGPT Marketing & Brand Strategy Prompts (Not the SWOT-Analysis Kind)
You asked ChatGPT for a marketing strategy and got a SWOT analysis, four bullets about "leveraging digital channels," and a competitive advantage that could describe any company in your industry. Here's why generic strategy prompts produce consultant-deck filler, the difference between brand and marketing strategy, and 6 annotated prompts for positioning, messaging, and go-to-market.
How to Write AI Prompts That Actually Work
You typed a prompt, got mush, added "be more detailed," and got longer mush. The problem isn't your wording — it's that your prompt has no constraints to fail against. Here's the mechanism, and six annotated templates you can paste today.
I want better answers by giving the AI a role/persona, but 'act as an expert' feels too generic. Help me do it right. The task: [WHAT I NEED HELP WITH] The kind of expert I imagine: [e.g. a skeptical CFO, a senior copywriter, a patient tutor] Write me a strong role-based prompt that: defines the persona specifically (experience, perspective, what they care about, how they communicate), tells it how to apply that lens to my task, and sets the tone and depth I want. Explain why a specific persona beats a generic one, and when role prompting actually helps vs. when it's just decoration. Give me the finished prompt and one variation with a different, useful persona for the same task.