The Best Claude Prompts for Keyword Research

If you've already tried using Claude to find keywords and ended up with generic, unhelpful answers, you're not alone. Most people use Claude as if it were a glorified search engine — and then complain about the results. The truth is that Claude is an extraordinarily powerful tool for keyword research, but only if you know how to talk to it.

Claude is not a traditional SEO tool like Ahrefs or SEMrush. It doesn't have access to real-time search volumes. But what it can do — and what it does better than any other tool — is reason about semantics, understand search intent, identify thematic clusters, and help you build a coherent content strategy. Used the right way, it becomes your personal 24/7 SEO strategist.

In this article you'll find ready-to-use prompts, tested and optimized, to get the most out of Claude in your keyword research activities. You'll also learn the prompt engineering principles behind each example, so you can adapt them to your specific industry and write your own prompts from scratch.

Why Claude Is Effective for Keyword Research

Before we dive into the prompts, it's worth understanding why Claude works so well in this context.

Claude is trained on massive amounts of web text. This means it has internalized not just words, but also search patterns: how people formulate questions, which synonyms they use, how search intent evolves along the buying funnel. It won't give you volumes, sure, but it gives you something equally valuable: deep semantic understanding of a topic.

Moreover, Claude excels at structured reasoning. You can ask it to organize keywords by cluster, by intent, by perceived difficulty — and it will do so with a consistency that purely data-driven tools often don't achieve.

8 Ready-to-Use Claude Prompts for Keyword Research

Prompt 1 — Seed Keyword Generation for a New Project

You are an SEO expert with 10 years of experience. I'm launching a blog about [your site topic, e.g. "sports nutrition for amateur runners"].

Generate 30 seed keywords that could be used as a starting point for my research. Organize them into 5 main thematic categories. For each keyword, briefly indicate the probable search intent (informational, navigational, transactional, or commercial).

Don't include search volumes — you don't know them. Focus on semantic relevance and variety of intent.

Why it works: This prompt leverages role assignment (giving Claude a specific role), which significantly improves response quality in technical domains. It also specifies what NOT to do (invent volumes), which reduces hallucinations. The request for categorization forces Claude to reason in a structured way instead of producing a chaotic list.

What to expect: A well-organized output with keywords grouped by theme and annotated with intent. Ideal as a starting point before moving keywords into a tool like Ubersuggest or Google Keyword Planner.

Personalization tip: Replace "sports nutrition for amateur runners" with your specific niche. The more precise you are in describing your audience, the more relevant the keywords will be.

Prompt 2 — Search Intent Analysis

Analyze the following keywords from the perspective of search intent. For each one, tell me:
1. Type of intent (informational, commercial, transactional, navigational)
2. Where the user is in the funnel (awareness, consideration, decision)
3. What type of content they expect to find (blog article, product page, comparison, tutorial, etc.)
4. An estimate of difficulty in ranking based on semantic topic competitiveness

Keywords to analyze:
- [keyword 1]
- [keyword 2]
- [keyword 3]
- [keyword 4]
- [keyword 5]

Respond with a markdown table.

Why it works: Asking for output in markdown table format is an output formatting technique that makes results immediately usable. The numbered list of required information reduces ambiguity and guarantees complete responses for each keyword. You're essentially giving Claude a rubric to follow.

What to expect: A clean, scannable table you can copy directly into Notion, Airtable, or Excel. The difficulty column is qualitative, not quantitative — use it as guidance, not as absolute data.

Personalization tip: Add a column by asking for "Suggested editorial angle" to get content brief ideas included in the analysis.

Prompt 3 — Semantic Expansion of a Main Keyword

Take the main keyword "[insert your main keyword, e.g. 'memory supplements']". and help me expand it semantically.

Generate:
1. 10 long-tail variations with informational intent
2. 10 long-tail variations with commercial or transactional intent
3. 8 related questions that users might search for on Google (formulated as real users would, not formally)
4. 5 LSI keywords (Latent Semantic Indexing) that I should include in the article to demonstrate topical authority

For the questions, be inspired by Google's "People also ask" and "Searches related to" sections.

Why it works: This prompt taps into Claude's ability to think like a real user, not like an SEO. The reference to Google's "People also ask" sections is a contextual anchor that aligns the output with what actually happens in SERPs. Separating variations by intent makes the result immediately actionable for editorial planning.

What to expect: A complete arsenal of keywords and questions you can use to structure a pillar article or an entire content cluster. The questions in particular are gold for FAQ sections and featured snippets.

Personalization tip: Add your target audience to the prompt: "...as someone 35-50 years old would search for them, new to supplements". Demographic specificity significantly improves the language of generated questions.

Prompt 4 — Building a Keyword Cluster

I'm an SEO consultant working for an e-commerce store in [product category, e.g. "hiking gear"]. I need to build a complete keyword cluster around the main topic "[main topic, e.g. 'trekking backpacks']".

Build a cluster organized like this:
- 1 pillar keyword (high difficulty, high presumed volume, mixed commercial/informational intent)
- 5-7 support keywords for blog articles (informational intent, long-tail)
- 3-5 keywords for category or product pages (transactional intent)
- 2-3 keywords for comparisons (commercial intent)

For each keyword also suggest:
- The most suitable page type (blog article, category page, product page, landing page)
- The ideal H1 title

Present the result as a structured editorial plan.

Why it works: This is a prompt of role-playing applied to business. Specifying that you're an SEO consultant and providing e-commerce context helps Claude tailor recommendations to the right business model. The requested structure produces output that's already an operational plan, not just a list of keywords.

What to expect: Essentially a draft editorial plan you can present to a client or use internally as a roadmap. With a few minutes of refinement, this becomes a professional document.

Personalization tip: If you already have competitors in mind, add them to the prompt: "Consider that my main competitors are [competitor names]. Try to identify keywords where I might find less semantic competition."

Prompt 5 — Competitor Keyword Analysis

Imagine you're a senior SEO analyst. I'm providing you with a list of the top-performing articles from a competitor in the [industry] sector:

[Paste here the competitor's article titles or URLs]

Based on these titles/URLs:
1. What are the main keywords the competitor is likely targeting?
2. What topic clusters is it trying to cover?
3. What semantic gaps do you see — in other words, what related topics does it NOT seem to be covering, which could represent opportunities for me?
4. Suggest 5 original editorial angles I could use to differentiate myself from this competitor.

Be specific and pragmatic. Avoid generalizations.

Why it works: Claude excels at textual analysis and pattern recognition. By feeding it competitor titles, you're providing concrete data to reason about. The request for "semantic gaps" is the most valuable part: Claude can identify what's missing from a content strategy with surprising precision.

What to expect: A qualitative analysis of the competitor's positioning and a list of opportunities. It doesn't replace an Ahrefs analysis, but complements it intelligently, especially in the creative phase.

Personalization tip: You can also paste meta descriptions along with titles to give Claude more context about the competitor's content type.

Prompt 6 — Keyword Research for Local Content

I need to create a local SEO strategy for [type of business, e.g. "a dental practice"] in [city or geographic area, e.g. "Bologna and surrounding area"].

Generate:
1. 15 keywords with explicit local intent (include city name or neighborhoods)
2. 10 keywords with implicit local intent (searches a local user would make without specifying the city)
3. 5 keywords for Google My Business and "near me" searches
4. The 5 most common questions a local customer might search for on Google before choosing [type of business]

For each group, indicate the recommended page type to target it.

Why it works: Local SEO has different logic from national SEO. This prompt explicitly distinguishes between explicit and implicit local intent — a distinction many SEOs overlook and that Claude understands well thanks to its semantic comprehension. The Google My Business reference aligns recommendations with actual mobile user behavior.

What to expect: A list of keywords usable for both website and Google My Business profile optimization. The frequently asked questions are perfect for FAQ sections or GMB posts.

Personalization tip: Add details about your target audience: "My ideal patients are 40-65 years old and primarily seek cosmetic and prosthetic treatments."

Prompt 7 — Editorial Brief from a Keyword

Create a complete editorial brief for an article optimized for the main keyword: "[insert keyword]"

The brief must include:
1. **Keyword analysis**: search intent, target user type, funnel stage
2. **Recommended structure**: H1, main H2s, H3s with notes on what to cover in each section
3. **Secondary and LSI keywords** to include in the article (at least 8-10)
4. **Questions to answer** in the text (with a view to People Also Ask)
5. **Differentiating angle**: what could make this article better than the average current results
6. **Recommended length** and ideal format (text only, with images, with tables, with video, etc.)

Write the brief as if it were intended for a copywriter with no SEO knowledge.

Why it works: This prompt transforms Claude from "keyword generator" to "content strategist". The final note — "as if intended for a copywriter with no SEO knowledge" — is a powerful prompt engineering detail: it forces Claude to be explicit and not take anything for granted, resulting in much more complete and actionable briefs.

What to expect: A 400-600 word document you can deliver directly to a copywriter or use as a personal outline. It's one of the highest-ROI prompts on this list.

Personalization tip: Add "The article tone should be [formal/conversational/technical] because the audience is [audience description]" to align the brief with your brand voice.

Prompt 8 — Identifying Low-Competition Keywords

I'm a blogger in the [industry] sector with a relatively new site (low Domain Authority, few backlinks).

Based on your knowledge of the sector, suggest 20 keywords that:
1. Have high specificity (long-tail, at least 3-4 words)
2. Answer very specific or niche questions that large generalist sites tend to overlook
3. Have predominantly informational intent
4. Are related to [specific blog topic, e.g. "vegan cooking for athletes"]

For each keyword briefly explain why you think it might have low competition.

Note: don't make up volume or difficulty data. Reason based on semantic specificity and the likelihood that large publishers haven't covered that micro-topic.

Why it works: This prompt is honest about Claude's limitations (it doesn't know actual difficulty data) but intelligently leverages its strength: semantic reasoning. Asking for keywords that "large generalist sites tend to overlook" is a creative framing that often yields genuinely interesting hidden gems.

What to expect: A list of highly specific long-tail keywords with reasoning behind each one. While the competition assessment is qualitative, it's often surprisingly accurate based on semantic analysis.

Personalization tip: Add your niche level: "My blog is specifically for [very specific audience], not the general public." This helps Claude identify keywords that are truly underserved in your micro-niche.