Unique Agency Logo
Hero Background

Keyword research is one of the most important parts of search engine optimization (SEO). It helps businesses understand what their audiences are searching for, identify valuable content opportunities, and attract relevant organic traffic to their websites.

However, relying solely on traditional keyword research tools can produce similar results across competing websites. When businesses use the same tools to research the same topics, they often end up targeting overlapping keyword lists.

Combining Claude, Anthropic’s AI assistant, with Semrush, an SEO and competitive research platform, offers a more contextual approach to keyword research. By combining AI-powered analysis with actual search data and business information, marketers can develop keyword strategies that better reflect customer needs and business opportunities.

The process becomes even more useful when it incorporates Google Search Console data, customer reviews, Reddit discussions, and sales call transcripts. These sources can reveal customer questions, pain points, and search-related topics that traditional keyword reports may not fully capture.

In this guide, you will learn how to set up a Claude project for keyword research, provide relevant business context, use Semrush and Google Search Console data, discover customer-driven keyword ideas, and turn your findings into an actionable SEO strategy.

Why Combine Claude and Semrush for Keyword Research?

Each tool serves a different purpose in the keyword research process.

Claude can analyze text, interpret customer feedback, organize topics, identify search-related questions, and generate keyword ideas based on your business context.

Semrush provides keyword and competitor research data, including estimated search volume, keyword difficulty, potential search intent, and the terms competing websites rank for, depending on the tools and reports available.

Google Search Console reveals the queries through which your website already appears in Google Search, along with impressions, clicks, click-through rate, and average position.

Combining these sources creates a more comprehensive research workflow:

  • •Use Claude to understand your business and generate relevant keyword ideas.
  • •Use Semrush to validate keyword opportunities with available search data.
  • •Use Google Search Console to identify optimization opportunities for existing pages.
  • •Analyze customer feedback to uncover the language people use when describing their needs and problems.

Claude should not be treated as a replacement for specialized keyword data tools. Search volume, keyword difficulty, and ranking information should be verified using reliable sources rather than assumed from AI-generated suggestions.

Step 1: Create a Dedicated Keyword Research Project in Claude

Instead of starting a new conversation for every keyword research task, create a dedicated project in Claude, provided the Projects feature is available in your account.

A project allows you to keep business context, instructions, and reference files organized in one place. This reduces repetitive explanations and helps maintain a consistent research process across multiple tasks.

How to Create the Project

  1. Open Claude and navigate to Projects.
  2. Create a new project.
  3. Give it a descriptive name, such as Keyword Research Workflow.
  4. Add a description explaining that the project focuses on keyword research and SEO opportunity analysis.
  5. Add project instructions and upload the reference files needed for future research.

You can use the project for several tasks, including competitor keyword gap analysis, Google Search Console opportunity discovery, competitor review analysis, and content topic research.

Configure Your Project Instructions

Project instructions define how Claude should approach research, which data sources it should use, and how it should present its findings.

You can use the following template and adapt it to your needs.

Ready-to-use project instructions:

Help me conduct keyword research and identify SEO opportunities using available Semrush data, uploaded Google Search Console exports, existing website URLs, customer reviews, competitor content, Reddit discussions, sales call transcripts, and the business context provided in this project.

For each task, follow the workflow appropriate to the requested analysis:

  1. Understand the business, its target audience, and its services before recommending keywords.
  2. Analyze keyword gaps between my website and competitors when the necessary data is available.
  3. Analyze Google Search Console exports to identify queries with optimization potential, including opportunities to increase clicks and improve existing pages.
  4. Extract keyword ideas from customer reviews, discussions, and sales transcripts while preserving the original meaning of customer statements.
  5. Validate suggested keywords using available Semrush data, including search volume, keyword difficulty, and potential search intent where available.
  6. Group keywords by search intent, customer journey stage, topic, and the most appropriate target page.
  7. Review existing website pages before recommending new ones to reduce content duplication and keyword cannibalization.
  8. Never invent search volumes, difficulty scores, or ranking data. Clearly identify missing information.
  9. Present findings in a structured table or CSV file when appropriate, including the keyword, idea source, available metrics, and recommended action.
  10. Follow the requested number of results and avoid unnecessary exploratory queries.

When a keyword cannot be validated because sufficient data is unavailable, separate it from ideas that have been verified.

These instructions help standardize the workflow, but they do not guarantee data accuracy. Results depend on the connected services, uploaded files, and available permissions.

Step 2: Add Your Business Context to Claude

One common keyword research mistake is analyzing search terms without providing enough information about the business.

A keyword may have substantial search demand but still be irrelevant to your services, target audience, or commercial objectives.

To avoid this problem, create a document called Business Context and add it to your Claude project.

The document should cover the following areas.

1. Products and Services

Describe exactly what your business offers, including its main services and any additional services you want to promote.

For an SEO agency, these might include technical SEO, content optimization, link building, local SEO, and e-commerce SEO.

2. Target Audience

Define the people or organizations you want to reach, such as small business owners, marketing managers, e-commerce businesses, or companies operating in a specific geographic market.

The more accurately you describe your audience, the easier it becomes to identify keywords that reflect its actual needs.

3. Customer Problems

Document the problems that lead customers to seek your services.

Examples include declining organic traffic, low conversion rates, slow website performance, or poor visibility in local search results.

These problems can inspire service pages, educational articles, comparison content, and frequently asked questions.

4. Main Competitors

List competitor names and website URLs. Distinguish between businesses that compete directly for your customers and websites that compete with you in organic search results.

A direct business competitor is not necessarily your competitor for every keyword.

5. Business Rules and Constraints

Specify relevant requirements, such as target markets, languages, services you do not offer, customer segments you want to attract, and claims that must not be made without supporting evidence.

Tip: Upload a list of your existing website URLs, ideally including each page title, page type, and target topic. Do not rely on URLs alone to understand page content. Include titles, descriptions, or crawl data whenever possible.

This information helps Claude recommend improvements to existing pages instead of creating new pages that target the same search intent.

Step 3: Connect Semrush and Prepare Google Search Console Data

Once your project is configured, prepare the data sources you will use to validate keyword ideas.

Connect Semrush Through MCP

If a Semrush connector is available in your Claude environment, follow the connection instructions provided in the connector interface.

Connection methods and available functionality may vary depending on your account and service configuration. Verify the connector's permissions and accessible data before beginning your analysis.

Depending on its supported features, a connection may allow you to retrieve keyword and competitor information without manually transferring every report.

If direct integration is unavailable, export the relevant Semrush reports as CSV files and upload them to Claude for analysis.

Export Google Search Console Data

Google Search Console helps you understand how your website performs in Google Search and is particularly useful for identifying optimization opportunities.

To export query data:

  1. Open Google Search Console and select your website property.
  2. Navigate to the Performance report.
  3. Open Search results.
  4. Choose an appropriate date range.
  5. Review clicks, impressions, click-through rate, and average position.
  6. Use the export option to download the data in a supported format, such as CSV or a spreadsheet, and upload the file to Claude.

For an established website, begin with a period long enough to identify meaningful trends. When useful, compare equivalent periods, such as the latest three months against the previous three months.

Keep in mind that Search Console may not show every query. Average position is also not a fixed ranking that every user sees, because search results can vary according to context and other factors.

Step 4: Conduct a Competitor Keyword Gap Analysis

Keyword gap analysis helps you discover search queries for which competitors appear in organic search while your website does not, as well as opportunities where your pages may have room to improve.

How to Perform the Analysis

  1. Identify competitors targeting the same market or topics.
  2. Collect keyword data for your website and competitor websites using Semrush.
  3. Upload the reports to Claude or use a supported direct connection.
  4. Ask Claude to compare competitor keywords with the queries for which your website appears.
  5. Group the opportunities according to service relevance, commercial intent, search volume, keyword difficulty, and existing page coverage.

Example of Opportunity Classification

Opportunity type

Recommended action

A relevant keyword with no suitable existing page

Evaluate creating a new page

A keyword targeted by an existing page with weak performance

Improve the existing page

High impressions but relatively few clicks

Review the title, description, and search intent alignment

A competitor keyword unrelated to your services

Exclude it or defer it

Several similar keywords with the same search intent

Consider targeting them on one comprehensive page

Not every competitor keyword represents a worthwhile opportunity. Evaluate its commercial value, relevance to your services, and realistic competitiveness before adding it to your strategy.

Step 5: Find Keyword Opportunities in Google Search Console

Your website may already be close to achieving better results for certain queries, but the relevant pages may need improvements to their content, titles, or overall usefulness.

Claude can help analyze Search Console exports and organize potential optimization opportunities by expected impact.

Ask it to identify:

  • •Queries with substantial impressions relative to clicks.
  • •Keywords for which the website appears in positions that may offer room for improvement.
  • •Pages that have lost clicks or impressions compared with a previous period.
  • •Queries related to existing pages but not adequately covered by their content.
  • •Search queries for which the website lacks a clearly relevant landing page.

Practical Example

Suppose a service page receives significant impressions for a particular query but has a low click-through rate.

Do not automatically create a new page. First, investigate:

  1. Does the title accurately communicate the page's purpose and match search intent?
  2. Does the meta description explain the page's value?
  3. Does the content answer the user's question?
  4. Is this the most appropriate page to appear for the query?
  5. Are other pages on the website competing for the same topic?

Then choose the appropriate action and measure the results using subsequent performance data.

Step 6: Turn Customer Reviews and Reddit Discussions into Keyword Ideas

SEO tools provide valuable data, but they do not always reveal how customers describe their problems or the objections that prevent them from buying.

Claude can analyze text from relevant customer interactions to uncover these insights.

Analyze Competitor Reviews

Collect publicly available reviews from relevant sources while respecting platform terms and privacy requirements.

Ask Claude to identify:

  • •Problems customers mention repeatedly.
  • •Services or outcomes customers praise.
  • •Reasons for dissatisfaction with previous providers.
  • •Questions that were not answered clearly.
  • •Phrases customers use to describe their needs.

Turn these themes into keyword ideas, then validate their relevance and available metrics in Semrush.

Analyze Reddit Discussions

Relevant Reddit discussions can reveal recurring questions, natural language patterns, and problems that people discuss in their own words.

Focus on conversations closely related to your services. Extract repeated questions, objections, and pain points that may inspire useful content.

However, a discussion is not proof of keyword search volume or market demand. Use it as a source of ideas and audience insight, then validate opportunities with search data where available.

Analyze Sales Call Transcripts

If you have sales call transcripts collected appropriately, Claude can help identify:

  • •Questions customers ask before purchasing.
  • •Concerns about price, expected results, and timelines.
  • •Reasons customers choose one service over another.
  • •Words customers use to describe their problems.
  • •Objections that could be addressed through educational content.

Remove personal and sensitive information before uploading transcripts to an AI service, and follow applicable company policies and legal requirements.

Validate Ideas Before Adding Them to Your Strategy

Do not automatically treat every idea from reviews, Reddit, or sales conversations as a keyword with confirmed search demand.

Use Semrush or another reliable data source to verify search volume, keyword difficulty, and potential search intent when those metrics are available.

Even if a term has no clear search volume, it may still be useful as a supporting topic or frequently asked question, provided it is relevant to your audience.

Step 7: Classify Keywords by Search Intent

Collecting keywords is only part of the process. You must also understand what users want to accomplish when they search.

Keywords can generally be grouped into four major intent categories.

1. Informational Intent

The user wants to learn something or find an answer.

Examples:

  • •What is search engine optimization?
  • •How do I choose keywords?
  • •Why is my website not appearing on Google?

These keywords are often suitable for educational articles, tutorials, and comprehensive guides.

2. Commercial Investigation

The user is comparing options before making a decision.

Examples:

  • •Best SEO tools.
  • •Comparison of keyword research tools.
  • •Best SEO agency for small businesses.

Comparison pages, reviews, and solution-focused content may be appropriate for these queries.

3. Transactional Intent

The user wants to purchase a service or take a specific action.

Examples:

  • •SEO service pricing.
  • •Request an SEO website audit.
  • •SEO company in Jeddah.

These keywords are generally suitable for service pages, pricing pages, and contact or lead-generation forms.

4. Navigational Intent

The user wants to reach a specific website or brand.

Examples:

  • •Semrush login.
  • •Official Claude website.

Evaluate these terms according to their relevance to your business. Avoid misleading targeting of other brands.

Step 8: Prioritize Keywords Based on Business Value

Once your keyword ideas have been collected and validated, create a structured spreadsheet to determine which opportunities deserve attention first.

Your spreadsheet can include the following fields:

Field

Purpose

Keyword

The search term being evaluated

Main topic

The topic cluster it belongs to

Search intent

Informational, commercial, transactional, or navigational

Search volume

The estimate reported by the source for the available period

Keyword difficulty

The competition estimate provided by the selected tool

Existing page

The URL currently targeting the keyword

Opportunity type

Create a page, improve a page, or update content

Priority

Critical, high, medium, or low

Idea source

Semrush, Search Console, reviews, Reddit, or another source

Next action

The specific task to implement

Do not prioritize keywords based on search volume alone. A lower-volume keyword may be more valuable when it is closely related to a service you provide and indicates a strong intention to become a customer.

Similarly, do not confuse a tool's keyword difficulty score with the actual difficulty of ranking your particular website. Domain strength, content quality, search intent, and the types of pages ranking in the results should also be considered.

Step 9: Export Your Keyword Research to CSV

After completing the analysis, ask Claude to organize the results into a CSV file that can be opened in Excel or Google Sheets.

The objective is to create an actionable research document rather than a long, unorganized keyword list.

Use the following prompt:

Keyword research CSV export prompt:

Create a CSV file containing validated keywords with the following columns: keyword, language, target market, search intent, main topic, search volume, keyword difficulty, data source, target URL, opportunity type, priority, and recommended action.

Separate validated keywords from preliminary ideas that lack sufficient supporting data. Do not invent search volume or keyword difficulty values. Leave unavailable fields blank, and include the data source and reporting period whenever known.

If a specific number of keywords is requested, return exactly that number without adding unnecessary rows.

Before finalizing the file, check for duplicate keywords, incorrect URLs, inconsistent metrics, and keywords that do not match the business's actual services.

Step 10: Turn Keyword Research into an Actionable Content Plan

The real value of keyword research lies in the decisions you make from the results, not in the final spreadsheet itself.

After grouping your keywords, assign each group to the most appropriate content type.

  • •Service pages: Target queries indicating that users are looking to hire a provider.
  • •Educational articles: Answer questions that require explanations or practical instructions.
  • •Comparison pages: Address queries involving comparisons between tools, products, or solutions.
  • •FAQ sections: Answer specific questions that customers frequently ask.
  • •Existing page updates: Improve relevant pages when updating them is more appropriate than creating new content.

Build a content calendar that specifies the target URL, primary keyword, supporting terms, search intent, business objective, publication or update date, and performance metrics.

Avoid creating a separate page for every similar keyword. Where appropriate, group closely related keywords with the same search intent into one comprehensive page to reduce duplication and internal competition.

Best Practices for Efficient and Accurate Research

When using Claude with Semrush, costs and usage limits may depend on file sizes, query volume, and the connection method.

To improve efficiency:

  1. Define the market, language, and research scope before starting.
  2. Upload only the files relevant to the current task.
  3. Request a specific number of results when a comprehensive list is unnecessary.
  4. Break large research projects into clear stages.
  5. Avoid repeating queries when the required data is already available.
  6. Verify the source of each metric before using it in reports.
  7. Maintain an organized record of validated keywords to avoid unnecessary duplicate research.

Review Claude's output before publishing content or committing to an SEO strategy, particularly when the recommendations affect business decisions or rely on data that may change over time.

Conclusion

Keyword research using Claude and Semrush combines contextual analysis, AI-generated ideas, and actual search data. Adding Google Search Console, customer reviews, and audience discussions provides a broader view of the opportunities that can support your SEO strategy.

Start by creating a dedicated Claude project, adding business context, collecting keyword data, analyzing competitors and existing pages, and validating ideas before prioritizing them.

Finally, turn your findings into an implementation plan that connects each keyword to an appropriate page, a clear objective, and measurable performance indicators.

This approach transforms keyword research from a simple list-building exercise into an ongoing process for understanding your audience, improving your content, and increasing your website's opportunities to appear in search results.