SEO writing has two small jobs that have huge consequences: choosing the right keywords and writing the meta description that earns the click. Both are perfect tasks for AI — short, structured, and pattern-driven. The trick is to organise keyword work around search intent, not raw volume, and to brief meta descriptions as ads, not summaries.
Most teams treat keyword research as a list-building exercise: feed a seed term to a tool, copy the highest-volume results, write articles around them. The reason this rarely works is that volume is a proxy — what really matters is whether the search intent matches what your business can deliver. AI is excellent at sorting keywords by intent if you ask it to. And meta descriptions, often left until the last second of the publishing process, are basically search-result ads — a separate copywriting task with its own rules.
Search queries map to four types of intent: navigational (the user wants a specific site), informational (the user wants to learn), commercial (the user is comparing options), and transactional (the user is ready to buy or sign up). Every keyword belongs to one type, and the type determines what kind of content can rank for it.
A common mistake is writing a blog post for a transactional keyword. "Buy noise cancelling headphones" is transactional — Google ranks product pages, not articles.
How to choose noise cancelling headphones
is informational — articles can rank. Same topic, two different intents, two different content types. AI is happy to classify intent if your prompt names the four buckets.
Weak prompt
Give me a list of 50 keywords about email marketing.
You will get 50 random phrases mixed across intents and difficulty levels. There is no plan, no grouping, no strategy. You still have to do the real work — sorting and prioritising — and AI just added noise rather than signal.
Use one prompt for intent-classified keyword research and a separate prompt for meta descriptions. Each does its job well.
Keyword research prompt
Act as an SEO strategist who has shipped content
plans for small B2B SaaS companies.
Topic area: email marketing automation for online
course creators.
Audience: Solo creators and small teams (under 5
employees) who already sell courses and want to
automate their email funnels.
Produce a keyword plan as follows:
1. 8 informational keywords (what the audience
would type to learn)
2. 6 commercial keywords (comparison, "best",
review-style)
3. 4 transactional keywords (close to buying)
4. 4 navigational keywords (brand-related, including
competitor names)
For each keyword, return:
- The keyword phrase
- Intent type (one of the four above)
- The content format that best matches the intent
(e.g. "long-form guide", "comparison post",
"landing page")
- One sentence describing the searcher's likely
state of mind
Avoid keywords with obviously high competition
(generic single-word terms). Prefer 3–5 word
long-tail phrases.
Meta description prompt
Act as an SEO copywriter.
Write 3 meta description variations for the
following blog post.
Title: "How to Build a 5-Email Welcome Sequence
for Your Online Course"
Target keyword: "welcome email sequence for online
courses"
Content summary: A practical guide to mapping out
the 5 emails a new student should receive in their
first week, with examples and templates.
Constraints:
- Between 140 and 155 characters
- Include the target keyword naturally
- End with a benefit, not a feature
- One variation must use a question
- One variation must use a specific number
- No words: "comprehensive", "ultimate",
"everything you need to know"
Return as a numbered list with character count
for each.
The keyword plan returns a structured grid you can drop straight into a content calendar. The meta description prompt returns three short variations — one factual, one question-led, one number-led — each within the character limit. You pick the strongest and ship.
Tip: After AI produces the keyword plan, paste it back and ask: "Group these keywords into 3 content clusters and propose a pillar page for each cluster." This is the start of a topic-cluster SEO strategy, in one follow-up prompt.
Pick a topic relevant to your work and run the keyword research prompt. Then ask AI to group the keywords into 3 content clusters with a pillar page idea for each. Save the output as the start of a real content plan.
Take three blog posts you have already published and ask AI to write 3 meta description variations for each. Compare to your current meta descriptions and pick the strongest version for each post.
Ask AI: "For each keyword in the plan, suggest one related question that people might also search." This gives you a ready-made FAQ section and helps you target featured snippets.
Sign in to join the discussion and post comments.
Sign inPrompt Engineering Projects & Real-World Applications
Twelve hands-on projects that turn prompt engineering theory into a portfolio. Build chatbots, content generators, RAG systems, and more.
Foundations of Prompt Engineering
The must-know basics of prompt engineering. Learn what prompts are, how AI models read them, and how to write clear instructions that get great results.
Prompt Engineering for Specific AI Tools
Tool-by-tool mastery — deep dives into ChatGPT, Claude, Gemini, GitHub Copilot, Midjourney, Stable Diffusion, and more. Learn the exact prompting techniques each platform rewards.
Prompt Engineering for Data Science & Analytics
Supercharge your data workflows with AI. 15 practical tutorials on using prompt engineering for data cleaning, EDA, machine learning, SQL, visualisation, and more.
Prompt Engineering for Developers
Use AI as your coding co-pilot. 18 tutorials on writing prompts to generate clean code, debug faster, write tests, build APIs, and ship better software.
Advanced Prompt Engineering Techniques
Master the powerful techniques AI experts use every day. Chain-of-thought, RAG, agents, function calling, prompt evaluation, and much more — 20 deep-dive tutorials.