GUIDE 03 – AI-POWERED SEO SERIES
Keyword Research & Clustering
with AI
Go from seed keywords to intent-mapped clusters with a prioritization score. Includes the SEO TAM Graph framework and the 4-factor scoring system used by enterprise brands.
REGULARLY UPDATED
INCLUDES FRAMEWORKS + PROMPTS
Why Traditional Keyword Research Is Broken
Most keyword research goes like this: pull 500 keywords from a tool, sort by volume, pick the highest-volume ones, and start writing content. This approach fails because it ignores the two things that actually matter – search intent and business relevance.
A keyword with 50,000 monthly searches means nothing if none of those searchers will buy from you. A keyword with 200 monthly searches could be a goldmine if every searcher has high purchase intent. The difference between average SEOs and the top 1% is not knowing more keywords. It is knowing which keywords to chase first and why.
Practitioner’s Note
“I am not interested in just looking at what keywords a business is targeting. I am interested in knowing what is the intent they are tapping into. Keywords tell you what people are typing. Intent tells you what they actually need. Once you identify that why, you unlock content ideas that totally differentiate you from competition.” – Apurv Singh, Dream SEO Masterclass Session 1
The Four Search Intent Types (Foundation)
Every search query falls into one of four intent categories. Before you do any keyword research, you need to understand these. They determine what content format to create, which pages to optimize, and how to prioritize your efforts.
| Intent Type | What the User Wants | Example Queries | Best Content Format | Revenue Impact |
|---|---|---|---|---|
| Informational | Learn more, get answers | “how to start a skincare routine,” “what is retinol” | Blog, guide, video, FAQ | Indirect (awareness) |
| Navigational | Go to a specific brand or page | “Nike official website,” “L’Oreal shampoo” | Homepage, brand page, store locator | High (brand intent) |
| Commercial | Compare before buying | “best laptops under $1000,” “vitamin C serum vs retinol” | Comparison blog, listicle, review page | High (mid-funnel) |
| Transactional | Ready to buy now | “buy running shoes online,” “vitamin C face serum for oily skin” | Product page, category page, landing page | Highest (bottom-funnel) |
Critical Shift to Watch
Informational searches are increasingly moving to AI chatbots and LLMs. If someone just wants to learn what retinol is, they will ask ChatGPT instead of clicking a blog link on Google. This means your informational content still matters for brand awareness and EEAT signals, but your commercial and transactional keywords are where revenue growth lives. Prioritize accordingly.
The SEO TAM Graph – Choosing the Right Keywords
This is a framework developed from working with enterprise brands across multiple geographies. It has helped win accounts from category leaders including major marketplaces in the UAE and Fortune 500 conglomerates. Every keyword you extract falls into one of four quadrants based on two variables: volume and relevance.
The Biggest Mistake SEOs Make
“I have seen fashion websites optimizing for keywords like ‘rangoli design.’ I have seen jewelry websites optimizing for ‘happy diwali’ and ‘infinity symbol.’ It makes no sense. You might get traffic, but you will not get sales. And worse, you will dilute your topical authority. The websites that got penalized – HubSpot, GeeksforGeeks – they lost traffic precisely because they tried to be everything for everyone instead of staying in their niche.” – Apurv Singh, Dream SEO Masterclass Session 1
Which Quadrant Should You Start With?
| Your Situation | Start With | Why |
|---|---|---|
| New website (under 6 months) | Low volume, high relevance | Low competition. You can rank faster. Even small traffic converts well because relevance is high. |
| Established site, want sales | High volume, high relevance | You have domain authority to compete. These keywords drive the most revenue at scale. |
| Limited content bandwidth | Low volume, high relevance | Fewer content pieces needed. Each piece targets a niche query with high conversion potential. |
| Already ranking for core terms | High volume, low relevance | Expand reach with a knowledge hub. Cast a wider net while core keywords continue driving revenue. |
| Client wants quick wins | Low volume, high relevance | Show results within 2-3 months. Big players ignore these keywords, giving you an opening. |
Step-by-Step: AI-Powered Keyword Research
Generate Seed Keywords with AI
Start by getting a broad list of keywords from AI. This works for any product or service. The key is providing enough context in your prompt – use the RCTCO format: Role, Context, Task, Constraint, Output.
Keyword Generation Prompt (RCTCO Format)
You are an SEO expert (Role). I sell {describe your product/service} targeted at {audience description} in {market/country} (Context). Generate a list of 50 keywords organized under informational, navigational, transactional, and commercial investigation intent (Task). Only include keywords relevant to my product category – no generic lifestyle keywords (Constraint). Present the output in a table with columns: keyword, search intent, estimated monthly volume, difficulty (1-100), and recommended content format (Output).
Accuracy note: AI-generated volume data is roughly 50-60% accurate compared to tools like Semrush or Google Keyword Planner. It is directionally useful for initial research. For precise data, validate with Google Keyword Planner (free) or Semrush (paid).
Validate with Google Keyword Planner
Take your AI-generated list to Google Keyword Planner (free inside Google Ads) for accurate volume data. If you have not run ads before, volumes will show as ranges (100-1,000 or 1,000-10,000). Once your first campaign runs for about 10-12 days, specific volumes unlock.
Alternative: If you have a Semrush or Ahrefs subscription, use their keyword explorer instead. The data is the most accurate available outside of Google and includes additional metrics like keyword difficulty and SERP features.
Map Keywords to the SEO TAM Graph
Take your validated keyword list and assign each keyword to one of the four quadrants. AI can do this for you – but critically evaluate the output. If you are presenting this to a founder and it is entirely AI-generated without review, they will catch it. They know their product and audience better than any model.
TAM Graph Mapping Prompt
Here is my keyword list with volumes:
— paste keyword + volume list —
My product is {describe product}. My target customer is {describe customer}. A keyword is “high relevance” if someone searching it has a strong probability of buying my product. A keyword is “low relevance” if they are seeking general information and unlikely to convert in the same session.
Please categorize each keyword into one of four quadrants: Low Volume/High Relevance, High Volume/High Relevance, Low Volume/Low Relevance, High Volume/Low Relevance. Present as a table and recommend which quadrant to prioritize first for a {new/established} website.
Apply the 4-Factor Prioritization Score
This is the scoring system used by enterprise brands managing 500+ keywords. It objectively decides which keywords to chase first by scoring each keyword across four factors on a scale of 1-5.
| Factor | How to Score (1-5 Scale) | Score 1 (Low) | Score 5 (High) |
|---|---|---|---|
| Volume | How many times per month is this searched? | Under 100/month | Over 10,000/month |
| Relevance | How closely does this match your product or service? | General, loosely related | Exact product/service match |
| Business Value | Will ranking for this keyword generate revenue? | Informational only, no conversion path | Directly tied to purchases |
| Competition | How hard is it to rank? (Inverted: low competition = high score) | Top 10 dominated by Amazon, Wikipedia, etc. | Few strong competitors, opportunity open |
The formula: Priority Score = Relevance + Business Value + Volume – Competition
Keywords with the highest priority score are your first targets. The formula naturally surfaces keywords where relevance and business value are high while competition is manageable.
Prioritization Prompt
You are an SEO expert. I am working on keyword prioritization using a four-factor scoring system: Volume (1-5), Relevance (1-5), Business Value (1-5), and Competition (1-5, where 5 = low competition). My product is {describe product/service}.
Please rate these keywords on all four factors, then calculate Priority Score using: Relevance + Business Value + Volume – Competition. Also recommend the funnel stage (top/middle/bottom) and a short note explaining your reasoning.
Keywords: {paste your keyword list}
Present as a table sorted by Priority Score (highest first).
Cluster Keywords into Content Groups
Once prioritized, group related keywords into clusters. Each cluster becomes one piece of content or one page. A single blog post should target a cluster of 5-15 related keywords, not just one keyword.
AI Prompt: “Group these keywords into topical clusters. Each cluster should represent one content piece. For each cluster, suggest: the primary keyword (highest priority score), supporting keywords, the content format (blog, product page, comparison, FAQ), a suggested title, and the target word count. Keywords: {paste your prioritized list}.”
Map Clusters to Pages (New vs. Existing)
The final step: decide whether each cluster needs a new page or can be addressed by optimizing an existing one. AI can cross-reference your cluster list against your current sitemap.
AI Prompt: “Here are my keyword clusters and here is my current sitemap {paste URLs}. For each cluster, tell me: (a) does an existing page already target this topic? If yes, which URL and what optimizations are needed? (b) If no existing page fits, recommend whether to create a new blog post, product page, or landing page, and suggest the URL structure.”
What AI Gets Wrong (And How to Catch It)
Volume data accuracy
AI estimates are 50-60% accurate. Always validate with Google Keyword Planner for critical decisions. The more location-specific or niche the keyword, the less accurate AI becomes.
Relevance misclassification
AI does not fully understand your business nuances. It might classify “protein powder for women” as informational when for your brand it is highly transactional. Always review the relevance column.
Difficulty score inflation
AI tends to overestimate difficulty for niche keywords. A keyword AI rates as “hard” might actually be easy if the top 10 results are weak, thin, or outdated content.
Missing long-tail opportunities
AI generates popular keywords well but misses hyper-specific long-tail queries your customers actually use. Supplement with Google Search Console query data and People Also Ask research.
The Critical Thinking Rule
“If you are presenting keyword data to a founder and it is entirely ChatGPT-generated, they might catch it. Founders know their business and product inside out. Even if you use AI, always critically evaluate the output. Every line item. That is how you build credibility and differentiate yourself from 99% of SEOs who dump tool-generated spreadsheets without thinking.” – Apurv Singh, Dream SEO Masterclass Session 2
Want the Full SEO Strategy System?
This guide covers keyword research and prioritization. For the complete 5-pillar framework including the SEO TAM Graph deep-dive, intent mapping workshops, content format selection, and the prioritization spreadsheet, explore the Dream SEO Masterclass.
Frequently Asked Questions
Do I need paid tools for keyword research?
No. ChatGPT or Claude can generate keyword lists for free. Google Keyword Planner (inside Google Ads) provides accurate volume data for free. Paid tools like Semrush ($129+/month) and Ahrefs add competitor analysis and historical data, but are not required to build a solid keyword strategy.
How many keywords should I target?
Start with 50 keywords for a small site, up to 500 for larger e-commerce or SaaS sites. The absolute number matters less than the quality of your prioritization. It is better to deeply optimize for 20 high-priority keywords than to spread thin across 200.
Is Semrush or Ahrefs better for keyword research?
Both are excellent and largely interchangeable. Semrush has a slightly friendlier UI and broader marketing toolkit. Ahrefs has stronger backlink data. The moment one launches a feature, the other follows within weeks. Pick the one that fits your budget and stick with it.
Should I target keywords for ChatGPT and AI search?
Not separately. There is no separate set of “AI keywords.” AI search engines pull from the same web content Google indexes. Strong SEO with proper structured data, EEAT signals, and clear content automatically makes you visible to AI platforms. Focus on building the best content for your target queries and AI citation follows.
How often should I redo keyword research?
Do a comprehensive keyword research round every 6 months. Between those rounds, monitor Google Search Console weekly for new queries that are bringing impressions but not clicks (page 2 opportunities). Seasonal businesses should do an additional round before each peak season.
Apurv Singh
Growth Architect – HQ Digital
12+ years in digital marketing. Built SEO for a global top-10 traffic website and multiple marketplace platforms. Currently consulting for brands across India, UAE, US, and Europe – including Fortune 500 conglomerates, Reliance Brands, and D2C companies in fashion, jewelry, health, and real estate. TEDx speaker. 300K+ followers across Instagram and YouTube.