Somewhere right now, someone is typing a question into ChatGPT instead of Google. They’re not searching for a “keyword” the way we’ve trained ourselves to think about it for the last 20 years — they’re asking a question the way they’d ask a knowledgeable friend. “What’s the best CRM for a 10-person sales team?” “How do I fix a slow WordPress site without hiring a developer?” “Is X better than Y for small business accounting?”
If your content strategy is still built entirely around traditional keyword research, you’re optimizing for half the picture. The other half — the one that determines whether AI platforms like ChatGPT, Perplexity, and Google’s AI Overviews mention your brand at all — requires a different kind of research entirely. That’s what AEO (Answer Engine Optimization) keyword and prompt research is about.
This guide walks through a complete, practical workflow for doing both: building a keyword list that still works for classic SEO, and layering on the prompt research that AEO requires. By the end, you’ll have a repeatable system, not just a list of tips.
Why Keyword Research and Prompt Research Aren’t the Same Thing
Traditional SEO keyword research revolves around search volume, keyword difficulty, and ranking on a results page. You’re targeting a query, and success looks like a blue link in position one through ten.
AEO doesn’t work that way. AI platforms don’t return ten ranked links — they synthesize an answer from multiple sources and decide, algorithmically, which brands and pages are worth citing or recommending. That means:
- Search volume matters less. A prompt might get typed into ChatGPT a thousand different ways, and none of those variations show up in a traditional keyword tool with meaningful volume.
- Intent and phrasing matter more. AI platforms respond to natural, conversational phrasing — often multi-sentence questions, comparisons, and follow-up prompts — not the clipped, fragment-style queries we’re used to typing into Google.
- Being “cited” replaces “ranking.” Your goal shifts from ranking for a keyword to being the source an AI model pulls from, quotes, or recommends when answering a related prompt.
None of this means traditional keyword research is obsolete — it’s still the foundation. AEO research builds on top of it.
Step 1: Build a Scalable Seed List
Every good research workflow starts with seeds — short, broad terms that describe your product, service, or niche. Don’t overthink this step. If you sell project management software, your seeds might be “project management,” “task tracking,” or “team collaboration tool.”
The trick isn’t picking clever seeds — it’s using modifiers to expand them into dozens or hundreds of directional ideas. Some of the most reliable modifiers across almost any niche include:
- Comparison modifiers: vs, alternative, compared to
- Evaluative modifiers: best, top, review
- Informational modifiers: how, why, what, guide, examples
- Commercial modifiers: pricing, cost, free trial
- Format modifiers: tool, software, template, checklist
Combine each modifier with your seed terms and you’ll walk away with a list that’s both broad enough to catch AEO-friendly conversational prompts and specific enough to feed into traditional keyword tools.
Step 2: Expand with Keywords Explorer and Matching Terms
Once you have a seed-and-modifier list, the next move is expansion. Tools like Ahrefs’ Keywords Explorer are built for exactly this: plug in a seed term and use a “Matching Terms” style report to surface every related query the tool has indexed, sorted by volume, difficulty, and other metrics.
This is where a lot of people get overwhelmed, because the output can run into the thousands. The fix is filtering aggressively:
- Use an include filter to keep only results containing specific words relevant to your business (your product category, your competitors’ names, or key modifiers like “best” or “vs”).
- Sort by relevance before volume. A 40-search/month keyword that perfectly matches buyer intent is often more valuable than a 4,000-search/month keyword that’s only tangentially related.
- Group results into topics, not individual keywords. AEO in particular rewards content that comprehensively covers a topic area rather than narrowly targeting one exact phrase.
Step 3: Vet Every Idea with the BID Method
Not every keyword or prompt idea deserves a piece of content. Some look promising on the surface but will never move the needle — either because they’re too competitive, too low-intent, or too disconnected from what you actually sell. A useful gut-check framework here is what’s sometimes called the BID method: evaluating ideas across Business value, Intent, and Difficulty before committing resources.
- Business value: Does ranking or being cited for this actually help you get customers, or is it just traffic for traffic’s sake?
- Intent: Is the person asking this question close to a buying decision, or are they in the very early stages of research?
- Difficulty: How hard will it realistically be to compete — either for a search ranking or an AI citation — given your current site authority and content depth?
Running every keyword idea through these three filters keeps you from chasing high-volume vanity terms that never convert and never get cited.
Step 4: Check the SERP Before You Commit
Even after filtering with BID, it’s worth spending 60 seconds looking at the actual search results page for your best keyword candidates. Two data points matter most:
- Referring domains of the top-ranking pages — a rough proxy for how much external validation those pages have earned.
- Domain Rating (or a similar authority metric) of the sites currently occupying page one.
If the top ten results are all high-authority, heavily-linked domains, that’s a signal the keyword will be a long game. If you spot a page ranking with a surprisingly low authority score, that’s often a sign there’s an opening — a realistic target you can compete for without years of link building.
Step 5: Where AEO Research Splits from SEO
This is the part most guides skip, and it’s the most important step if your goal is AI visibility specifically.
Map prompts at the topic level, not the keyword level
Conversational AI prompts can’t be targeted the same way individual keywords can. Nobody types the exact same 40-character query into ChatGPT that they’d type into Google. Instead of chasing exact-match prompts, map out the topic cluster a prompt belongs to, and build content that thoroughly answers every angle of that topic — background, comparisons, pricing, use cases, common objections.
Study how AI platforms actually phrase and answer questions
Spend time in ChatGPT, Perplexity, and Google’s AI Overviews typing in prompts related to your niche. Pay attention to:
- How the AI phrases its answer (tone, structure, length)
- Which sources it cites or references
- Whether it recommends specific brands, and which ones
This isn’t just competitive research — it’s a direct look at the exact “exam” your content needs to pass to get cited.
Find your AI mention gaps
One of the most actionable AEO tactics is identifying places where your competitors get mentioned by AI platforms and you don’t. If you have access to a brand-monitoring tool, look specifically at queries containing commercial-intent modifiers — “best,” “top,” “vs,” “review,” “alternative” — and filter for results where a competitor is cited but your brand is absent. That gap list becomes your AEO content priority list.
It’s also worth treating this as an ongoing process rather than a one-time audit. AI citation sources shift frequently as models get updated and re-index the web, so a mention gap you find today might look different in a few weeks. Revisit it on a regular cadence — monthly at minimum.
Prioritize content depth over content volume
AI platforms tend to synthesize answers from pages that comprehensively cover a topic, not from thin pages that barely mention it. A single deep, well-structured guide that answers a topic from five different angles will typically outperform five separate shallow posts, both for traditional SEO and for AEO citation potential.
Putting the Workflow Together
Here’s the full process, start to finish:
- Brainstorm seed terms tied to your product, service, or niche.
- Expand with modifiers (how, best, vs, guide, alternative, pricing) to generate a broad idea list.
- Feed seeds into a keyword research tool and use matching-terms style expansion to surface hundreds of related queries.
- Filter with an include filter to strip out irrelevant noise.
- Vet every surviving idea with BID — business value, intent, difficulty.
- Check the SERP for referring domains and authority signals on your shortlist.
- Layer in AEO-specific research — map topics instead of exact prompts, study how AI platforms phrase and cite answers, and hunt for AI mention gaps against competitors.
- Build comprehensive, topic-level content rather than chasing narrow, exact-match keywords.
- Monitor and repeat — both keyword rankings and AI citation gaps change over time, so treat this as a living process, not a one-off project.
Final Thought
The businesses winning in AI search right now aren’t the ones with the fanciest tools — they’re the ones treating keyword research and prompt research as two connected halves of the same strategy instead of two separate projects. Build your foundation with solid, filtered keyword research. Layer AEO-specific prompt mapping and citation-gap analysis on top. Revisit both regularly. That combination is what actually gets you found — whether someone’s typing into a search bar or asking an AI assistant a question out loud.