One Keyword Research Framework for Both Google and AI Search

One Keyword Research Framework for Both Google and AI Search

Most teams still run keyword research like it’s 2017:
export a big list from a tool, sort by volume, pick a few “easy” ones, and start writing.

That approach breaks down fast when your buyers aren’t just using Google. They’re also asking questions inside ChatGPT, Perplexity, Gemini, and other AI tools—and those questions don’t always show up neatly in keyword tools.

You don’t need two separate processes (“one for SEO, one for AEO”). You need one simple framework that works for both.

This is the one we like to use at Sagashi.

The goal: think in topics and questions, not just keywords

Before we get into steps, it helps to reset the goal.

Good keyword research should:

• reflect how your buyers actually think and ask questions
  
  
• give you a clear map of “what we should talk about and why”
  
  
• connect naturally to content ideas and formats
  
  
• work across Google results and AI answers
  
  

If you come out of research with a spreadsheet no one wants to look at, you don’t have a framework—you have a guilt document.

The 3×3 framework: Intent, Question Shape, Depth

Here’s the simple model we’ll use.

For each keyword or topic, we’ll look at three things:

1. Intent – What is the person trying to achieve?
   
   
2. Question Shape – How would they actually phrase it?
   
   
3. Depth – How deep of an answer do they need?
   
   

Do this, and you end up with content ideas that:

• match searcher intent for Google
  
  
• match natural language questions for AI tools
  
  
• match the right level of detail so people don’t bounce or feel overwhelmed
  
  

Let’s break those down.

1. Intent: what are they really trying to do?

Intent is the “why” behind a search.

At a basic level, most queries fall into a few buckets:

• Learn: “what is…”, “how does…”, “why does…”
  
  
• Compare: “best…”, “vs…”, “alternatives to…”
  
  
• Do / Buy: “pricing”, “template”, “book a demo”, “near me”
  
  

When you look at a keyword, ask:

• What would this person be doing 5 minutes before searching this?
  
  
• What are they hoping to do 5 minutes after?
  
  

Example:
“SEO vs AEO”

• Before: maybe they heard both terms in a webinar or tweet.
  
  
• After: they want to decide if they should change their strategy, hire someone, or ignore AEO.
  
  

Intent here is: compare and decide on a direction, not just “get a definition.”

Once you know the intent, you can:

• Design content that matches that job.
  
  
• Avoid writing fluffy explainers where you should be giving clear recommendations.
  
  

2. Question shape: how would they actually ask?

AI tools especially are driven by how people phrase their questions.

For each topic, write down a few natural variations that feel like real sentences, not just tool output.

Example topic: “answer engine optimization”

Possible question shapes:

• “What is answer engine optimization and how is it different from SEO?”
  
  
• “Do I really need AEO or is SEO enough?”
  
  
• “How do I optimize my content for AI answers?”
  
  

These may or may not show up as exact keywords with volume—but they reflect real conversation.

This helps you:

• Turn headings into natural questions.
  
  
• Write intros that sound like you’re responding to the reader.
  
  
• Seed FAQs with questions that AI tools will recognize as useful.
  
  

3. Depth: how deep of an answer do they really need?

Not every query deserves a 3,000‑word guide.
Not every query can be satisfied with a 2‑line FAQ.

For each topic, ask:

• Is this a quick question? (definition, simple how‑to)
  
  
• Is this a decision question? (needs pros/cons, context, comparison)
  
  
• Is this a strategic question? (needs frameworks, examples, nuance)
  
  

Examples:

• “What is AEO?” → quick question (can start with 2–3 sentences + optional deeper sections)
  
  
• “SEO vs AEO” → decision question (requires clear comparison and recommendations)
  
  
• “How should B2B SaaS think about search in an AI world?” → strategic question (needs frameworks, cases, and caveats)
  
  

Depth then informs:

• how long the main piece should be
  
  
• whether it should be one article or a series/cluster
  
  
• how many supporting assets you need (FAQs, checklists, templates)
  
  

Step-by-step: running the 3×3 framework

Now let’s put this into a repeatable process.

Step 1: Start with seed topics, not just keywords

Make a short list of:

• products/features you sell
  
  
• problems your buyers talk about
  
  
• outcomes they care about
  
  

For Sagashi, examples might be:

• “SEO for B2B SaaS”
  
  
• “answer engine optimization”
  
  
• “content systems for search”
  
  
• “AI search traffic”
  
  

These are your topics.

You can still use tools (Ahrefs, Semrush, etc.) to:

• discover related keywords
  
  
• see rough volume and difficulty
  
  
• find “people also ask” ideas
  
  

But you’re not letting tools dictate reality—you’re using them as input.

Step 2: Map intent for each topic

For every topic/keyword you care about, ask:

• Is this Learn / Compare / Do?
  
  
• What job is this query trying to get done?
  
  

You can do this quickly in a simple sheet:

• Column A: Topic / keyword
  
  
• Column B: Intent (Learn / Compare / Do)
  
  
• Column C: “Job to be done” in one sentence
  
  

Example row:

• Topic: “SEO vs AEO”
  
  
• Intent: Compare
  
  
• Job: “Decide how to update our search strategy for AI”
  
  

This alone will make your content 10x more focused.

Step 3: Add 3–5 natural question shapes

For each topic, write:

• 1–2 “what is…” or “how does…” versions
  
  
• 1–2 “how do I…” or scenario‑based versions (“for a small team”, “for SaaS”)
  
  

You can:

• pull some from “People also ask” sections
  
  
• borrow phrases from sales calls and emails
  
  
• make them up based on how you actually talk to customers
  
  

Drop them into a column (or separate tab if you want to keep things neat).

This set of questions becomes:

• headings in your content
  
  
• FAQ questions at the bottom
  
  
• ammunition for future content angles
  
  

Step 4: Decide on depth and format

For each topic, answer:

• Depth: Quick / Decision / Strategic
  
  
• Format:
  
  
  
  • Quick → FAQ entry or short explainer
    
    
  • Decision → comparison post, guide with a “recommendation” section
    
    
  • Strategic → longform guide, pillar page, or multi‑part series
    
    

Example:

Topic: “SEO vs AEO”

• Depth: Decision
  
  
• Format: comparison article with:
  
  
  
  • definitions,
    
    
  • pros/cons,
    
    
  • “when you need which”,
    
    
  • clear recommendation for each type of business.
    
    

Topic: “What is AEO?”

• Depth: Quick
  
  
• Format: definition page or section, with short follow‑up sections for people who want more.
  
  

Now your spreadsheet isn’t just a dump of keywords; it’s a content roadmap.

How this plays out for Google and AI

Using the same framework, you unlock wins across both worlds.

For Google

• Intent alignment → users stay longer and click deeper.
  
  
• Clear question shapes → better chance at matching long‑tail, natural‑language queries.
  
  
• Right depth/format → fewer bounces from mismatched pages.
  
  

You’re sending Google signals that:

• this page actually solves the user’s problem
  
  
• people don’t pogo‑stick back to the results
  
  
• the content is comprehensive enough for the query type
  
  

For AI tools and answer engines

• Clear questions and headings → easier to detect what each section is about.
  
  
• Direct answers at appropriate depth → easier to quote or summarize correctly.
  
  
• FAQs and structured content → more reusable bits for specific prompts.
  
  

You’re essentially telling AI:

• “Here’s the exact answer to that question.”
  
  
• “Here’s the context if you need it.”
  
  
• “Here’s how to break this into steps if someone asks for a how‑to.”
  
  

Same content, two layers of payoff.

Example: applying the framework to one topic

Let’s walk through a single example:
Topic = “SEO for B2B SaaS”

1. Intent
   
   
   
   • Mostly Learn + Strategic
     
     
   • Job: “Figure out what SEO should look like for a B2B SaaS product and whether it’s worth the effort.”
     
     
2. Question shapes
   
   
   
   • “How should B2B SaaS companies do SEO?”
     
     
   • “Is SEO worth it for early‑stage B2B SaaS?”
     
     
   • “What’s different about B2B SaaS SEO?”
     
     
   • “How long does SEO take to work for SaaS?”
     
     
3. Depth
   
   
   
   • Likely Strategic → needs frameworks, examples, caveats.
     
     
4. Format
   
   
   
   • Longform guide / pillar page.
     
     
   • Sections:
     
     
     
     • how B2B SaaS buying journeys work,
       
       
     • what that means for search,
       
       
     • key playbooks by stage,
       
       
     • realistic timelines,
       
       
     • how to combine SEO + AEO for SaaS.
       
       

Then, layer in AEO tactics:

• Clear definition section: “What we mean by B2B SaaS SEO”
  
  
• FAQ: “Is SEO worth it for pre‑product‑market fit SaaS?”
  
  
• Headings that mirror questions: “How long does B2B SaaS SEO take?”
  
  

Now you have a piece that:

• can rank on Google
  
  
• can answer specific AI prompts
  
  
• can serve your sales team as a resource
  
  

When you don’t need to overthink it

Not every single keyword needs a full 3×3 analysis.

You can be lighter when:

• the keyword is clearly navigational (“[brand] pricing”)
  
  
• the intent is obvious and low‑stakes
  
  
• the topic is a small supporting piece, not core to your strategy
  
  

In those cases, a quick gut check on intent + one or two natural question shapes is enough.

Save the full framework for:

• core problem/solution topics
  
  
• important “education” pieces
  
  
• high‑value comparison or “vs” content
  
  

Turning the framework into a living document

The worst fate for a keyword sheet is to be created once and never touched again.

To keep this useful:

• Revisit it monthly or quarterly.
  
  
• Add new questions you hear from sales/support.
  
  
• Mark topics as “planned”, “in progress”, “live”, and “needs refresh”.
  
  

And importantly:

• Connect each topic to outcomes where you can (traffic, signups, demos, revenue).
  
  
• Note which topics pull their weight and which ones are nice but not critical.
  
  

Over time, this becomes:

• your search roadmap
  
  
• your content roadmap
  
  
• and a shared source of truth across marketing, sales, and leadership
  
  

Not just “SEO stuff in a silo.”

How Sagashi uses this in practice

Internally, we treat this 3×3 framework as the spine of search strategy:

• Intent keeps us honest about why a piece exists.
  
  
• Question shape keeps us close to how people actually talk and search.
  
  
• Depth keeps us from under‑ or over‑serving a topic.
  
  

Then we layer SEO and AEO on top:

• SEO to make sure the right people can find it.
  
  
• AEO to make sure AI tools can understand and use it.
  
  

If you’d like, you can send over 5–10 of your “must win” topics, and we can show you how we’d fill out this framework for one of them. It’s usually enough to get the whole system clicking into place.