Keyword Research with AI

What You'll Need

Access to either Semrush or Ahrefs (we'll use Semrush in examples), a ChatGPT subscription for advanced analysis, and AlsoAsked for question research. You should have your primary topic or seed keyword identified, plus 30 minutes of uninterrupted time for the AI analysis phase. A spreadsheet tool for organizing results is essential.

Step 1: Seed Keyword Expansion

Time: 30 minutes | Tool: Semrush Start in Semrush's Keyword Magic Tool and enter your primary seed keyword. Set your target location and click "Search." Export the first 1,000 results to Excel, filtering for keywords with search volume above 100 and keyword difficulty below 70 for most sites. Next, use Semrush's "Questions" filter to isolate question-based keywords, then export these separately. Navigate to the "Related Keywords" section and export another 500 terms. This gives you roughly 1,500 raw keywords across informational and commercial intents. Create three separate tabs in your spreadsheet: "High Volume Keywords," "Question Keywords," and "Related Terms." This separation will be crucial for the AI analysis phase.

Step 2: Intent Classification with AI

Time: 45 minutes | Tool: ChatGPT Open ChatGPT and use this prompt: "I'm going to give you a list of keywords. For each keyword, classify the search intent as: Informational (INFO), Commercial Investigation (CI), Transactional (TRANS), or Navigational (NAV). Also rate the content difficulty from 1-10 where 10 requires deep expertise. Format as: Keyword | Intent | Difficulty | Content Angle Suggestion." Paste 50 keywords at a time from your "High Volume Keywords" tab. ChatGPT will provide intent classifications plus content angle suggestions you wouldn't get from traditional tools. Copy these results back to your spreadsheet, adding columns for Intent, Difficulty, and Content Angle. Repeat this process for your question keywords, but modify the prompt: "These are question-based keywords. Classify intent and suggest whether they'd work better as FAQ content, dedicated articles, or section headers within larger pieces."

Step 3: Question Research Deep Dive

Time: 30 minutes | Tool: AlsoAsked Enter your main seed keyword into AlsoAsked.com and set your target country. The tool will generate a visual map of "People Also Ask" questions from Google. Export the results as CSV - you'll typically get 100-200 question variations. Focus on questions that appear in multiple branches of the tree visualization, as these indicate strong user interest. Cross-reference these questions with your existing keyword list to identify gaps. Add any new question-based keywords to your "Question Keywords" tab. Use AlsoAsked's depth levels (the tool shows questions branching 3-4 levels deep) to understand question hierarchies. Level 1 questions are broader topics, while Level 3-4 questions reveal specific pain points perfect for detailed content sections.

Step 4: Semantic Clustering Analysis

Time: 45 minutes | Tool: ChatGPT Return to ChatGPT with this advanced prompt: "I'll provide keywords related to [your topic]. Group these into semantic clusters where keywords share similar user intent and could be targeted on the same page. Name each cluster and suggest the primary keyword for each group. Also identify any keyword cannibalization risks." Paste your complete keyword list in batches of 100. ChatGPT will identify semantic relationships that traditional clustering tools miss, grouping keywords by actual user intent rather than just text similarity. This reveals opportunities to target 15-20 related keywords on single optimized pages. Create a new spreadsheet tab called "Keyword Clusters" and organize the AI's suggestions into themed groups. Mark the primary keyword for each cluster and note which keywords should be secondary targets or semantic variations within the content.

Step 5: Competitive Intent Verification

Time: 30 minutes | Tool: Ahrefs Take your top 20 primary keywords from the clustering analysis and run them through Ahrefs' Keywords Explorer. For each keyword, click "SERP Overview" to analyze the top 10 results. Look for three key signals: content format patterns (listicles vs. guides), content depth (word counts), and commercial vs. informational results. Export the SERP analysis data and cross-reference it with your ChatGPT intent classifications. If ChatGPT classified a keyword as "Informational" but the SERP shows mostly commercial pages, flag this as a mixed-intent keyword requiring a hybrid content approach. Use Ahrefs' "Also rank for" feature for each of your target keywords to discover additional semantic terms that top-ranking pages successfully target together. Add these to your keyword clusters as supporting terms.

Common Pitfalls

• Relying solely on search volume without considering intent alignment with your business goals
• Grouping keywords by text similarity rather than actual user intent and content requirements
• Ignoring question-based keywords which often drive featured snippets and voice search results
• Not validating AI intent classifications against actual SERP results before building content strategies

Expected Results

You'll have a prioritized keyword database with 300-500 keywords organized into 15-25 semantic clusters, each with clear intent classifications and content angle suggestions. Expect to identify 50-100 question-based keywords perfect for FAQ sections and featured snippet opportunities. Your keyword list will include semantic variations that traditional tools miss, giving you a competitive advantage in content planning and a clear roadmap for the next 6-12 months of content creation.