Prompt engineering
80% of the result is in the correct query.
Prompt engineering is not magic or programming. This is the skill of explaining to a machine what you need, as clearly as you would explain to an intern. Only the trainee can ask again, and the neural network will give the result and wait for the next task.
4 tricks that always work
Tell the model who she is, what her situation is, what her limitations are. “You are a marketer with 9 years of experience in e-com. You are writing a post for a Telegram channel about AI tools. The audience is the owners of online stores. The tone is expert, without hype.”
The same task with different roles gives different results. Ask “write like a copywriter” - you will receive a selling text. Ask “write like an editor” and you will get dry and structured.
Instead of “make it beautiful,” show an example of what you think is beautiful. “Here are 3 examples of good posts (insert links). Write the fourth in the same style.”
Tell the model what format the answer is needed in: table, list, 3 paragraphs, 80 words. “Write 5 headline options in the format: [number] - [title] - [why it works].”
Try it yourself
Below is an AI practice tool. Try writing a query with context and role, and then without them. Compare the difference in answer quality.
AI answers are free, but limited by the overall site limit. If the limit is reached, come back later: repeated clicks will not speed up the response.
Feel the difference
The same request: on the left is a “naked” prompt, on the right is yours, with a role, context and format. Write yours and compare the answers live.
Task: Write a selling post on Telegram about the launch of a new online marketing course.
AI answers are free, but limited by the overall site limit. If the limit is reached, come back later: repeated clicks will not speed up the response.
Challenge: collect a strong prompt
Write a prompt for the task below. The AI will check it against 4 lesson criteria and suggest what to strengthen.
Task: We need a prompt that will force the neural network to write 3 versions of a post for the Telegram channel of a B2B company about a new AI analytics service.
- •Role assigned (who writes)
- •There is a context and target audience
- •Tone of voice indicated
- •The output format is specified (structure, volume)
AI answers are free, but limited by the overall site limit. If the limit is reached, come back later: repeated clicks will not speed up the response.
Prompt template for today
You are [role: marketer / copywriter / SMM specialist / editor] with [N years] experience in [niche]. You are writing [format: post / article / letter / briefing] for [channel: Telegram / Instagram / email] about [topic]. Target audience: [describe target audience]. Tone: [confident / cocky / expert / friendly]. Goal: [sales / engagement / information]. Volume: up to [N] words/characters. Task: [specific task].
Which technique has the greatest impact on the quality of the answer?
Why indicate the role in the prompt (“you are an e-com marketer”)?
What is more effective than the phrase “make it interesting”?