How To Find Competitor Keyword Gaps With AI (Free Method)

By Corey Frankosky · Last Updated August 17, 2026

If your competitors rank for terms you do not, rank with a better page type, or get cited in AI Overviews while you do not, you have a keyword gap. Most guides on this assume you are paying for Semrush or Ahrefs. You do not need to be. You can find competitive keyword gaps with the Google Keyword Planner, the free SpyFu plan, and Claude, and get a categorized list of what to build in about fifteen minutes. This walks through that exact workflow using a real example, then covers what changes if you do have paid tools.

1What a keyword gap actually is

A keyword gap is not just "they rank and I do not." There are five different kinds, and each one calls for a different fix. Knowing which one you are looking at saves you from writing a new article when you should have expanded an existing page.

This process works best when you already have an established site ranking for something. If you are starting from zero, you do not have gaps yet, you have a blank content plan. The gap analysis is for when you want to out-rank specific competitors.

Question to Answer:

Are you already ranking for some keywords, or are you building your first set of pages?

2Watch the full tutorial

This is a quick one. I run the whole process start to finish on a real site I had never used before, so you can see exactly what the exports and the Claude output look like.

How To Find Competitive Keyword Gaps Using Google Keyword Planner And AI

Question to Answer:

Do you have access to a Google Ads account, which is what you need to use Keyword Planner?

3Find your real SEO competitors with SpyFu

Your SEO competitors are not always your business competitors. The companies your sales team worries about may never show up in Google for your target searches, and the sites that do rank may sell something completely different, or nothing at all. A local HVAC company competes in search with HomeAdvisor, Angi, and Reddit threads as much as with the installer down the road.

The fastest way to find them: go to SpyFu, enter your domain, and look at the top organic competitors it returns. You do not need a paid plan for this. The free view gives you the competitor list, which is all you need for this step.

For the example I ran, I picked a random AI video generation platform I had never used, Powtoon, so nothing about the result was influenced by what I already knew. SpyFu returned competitors including Videoscribe, Animaker, and InVideo.

Which competitors to actually use. You want the biggest competitors closest to your business type. When you run them through the next step, pay attention to how many keyword ideas each one generates. In my example, Powtoon and Videoscribe came back with around 725 keyword ideas each. Animaker came back with nearly 2,800. That difference tells you Animaker is covering far more ground, which makes them the more useful comparison, because that is where the gaps are.

You can also cross-check by searching your top non-branded keywords in Google and noting which domains keep appearing. If the same three or four sites show up repeatedly, those are your search competitors for that topic cluster.

Question to Answer:

Who are the three or four domains that actually show up when you search your main non-branded keywords?

4Export keyword data from Google Keyword Planner

Open Google Keyword Planner from inside your Google Ads account. Then do this for your own domain and for each competitor:

1. Go to Discover new keywords.
2. Switch to the Start with a website tab.
3. Enter the domain and select Use the entire site.
4. Click Get results.
5. Download the results as a CSV.

Repeat for each competitor. Three or four files is plenty. Name them clearly so you know which is which when you upload them.

What you get in each export is the keyword, average monthly searches, competition, and the top of page bid range. That bid data is more useful than people realize. It tells you which keywords have real commercial value, because advertisers do not pay $15 a click for terms that do not convert.

One thing to notice while you are in here. You will spot keywords with real volume that describe features or tools you do not offer at all. In my example, "remove the background from an image" came back with a large monthly search volume. That is not a content gap. That is a product gap. If a competitor has a free tool driving that traffic and you do not, building the tool may be the actual answer rather than writing an article about it.

Question to Answer:

Do you have your own CSV plus at least two competitor CSVs downloaded?

5Upload the CSVs to Claude and ask for gaps

Now go to Claude and upload all of the CSV files at once. Your prompt does not need to be perfect. Something close to this works:

Asking for search volume and bid data matters. Without it you get a list of terms with no way to judge which are worth the effort.

What comes back is your gap list, with the keywords your competitors are considered relevant for and you are not. In my example that surfaced things like AI voice generator, text to speech voices, and a range of animation software and tools terms where Powtoon simply was not showing up.

Claude will usually give you more than you asked for. In my run it produced a full written explanation and breakdown I had not requested. That is generally a good problem to have here.

Question to Answer:

Did your gap list include search volume and bid data, or just keywords?

6The categories matter more than the keyword list

This is the part that makes the whole exercise worth doing, and it is the reason to run it through AI rather than sorting a spreadsheet by hand.

A raw gap analysis gives you 340 keywords. Nobody can act on 340 keywords. What you actually want is five categories.

In my example the gaps grouped into themes like AI voice and text to speech, animation software and tools, and a few others. That is something you can take to a client or a team and say: here are the five areas where we are losing ground, let's start with AI voice and text to speech, because there is a lot to build inside that one category and a lot of monthly search volume we are missing.

Then the decision inside each category is simple:

• If you have nothing in that category, build it. New pages, new tools, new content covering the theme.
• If you already have something, expand it. Powtoon does have text to speech. So the move there is not a new page, it is going back to Keyword Planner, pulling everything related to text to speech, seeing how people actually search for it, and building that out properly.

Worth saying plainly: the content only takes you so far. In a category like this, whoever builds the best text to speech tool is going to win the category eventually. Keyword gaps tell you where to compete. They do not substitute for having something worth ranking.

There is a quote I like on this from Boyuan Deng at RankAI: "Competitive analysis fails when spreadsheets have keywords, topics, and traffic data but no decisions." That is exactly the trap. The categories are what turn data into a decision.

Question to Answer:

Can you name the five categories where you are losing to competitors, without opening a spreadsheet?

7Level two: find page gaps using top pages by traffic

Once you have the keyword gaps, you can take this a step further and find page gaps. This part requires a paid keyword tool to pull top pages by organic traffic, and a paid Claude plan to handle the file sizes. For $20 a month, Claude is worth it for this alone.

Here is the process:

1. Export your top pages by organic traffic for your own domain. In my example this came back with about 1,200 rows.
2. Export the same report for each competitor. You are looking at which of their pages actually pull traffic and what those pages rank for.
3. Upload all of them to Claude together with your own.
4. Ask: "These are competitors' top pages plus my own pages. Give me 25 high-impact pages I should create."

The output is a build list rather than a keyword list. In my run it returned specific pages: AI voice generator, AI subtitle generator, AI cartoon maker, AI presentation maker, and comparison pages like "[competitor] alternative" for each major player in the space.

You also get the traffic evidence behind each one. For instance, Animaker was pulling roughly 364 clicks a month from a single page targeting "AI text animation maker," which Powtoon had nothing for. That is a page you can point at and justify.

Do not skip the alternative pages. "[Competitor] alternative" searches come from people already shopping, already unhappy with something, and already aware of the category. They are some of the highest-intent traffic in any market.

Question to Answer:

Which single competitor page drives the most organic traffic in your category, and do you have anything comparable?

8Where Semrush and Ahrefs fit in

If you already pay for Semrush or Ahrefs, they do this faster and with more precision. If you do not, the workflow above gets you most of the way there for free.

Semrush Keyword Gap. Add your domain and your competitor domains. Two filters carry most of the value. Missing shows terms competitors rank for and you do not. Weak shows terms where you rank but they rank higher. Sort both by volume and difficulty, then filter for intent that actually matches the page type you would build.

Ahrefs Content Gap. Use this to confirm what Semrush surfaced. The overlap view shows you how many competitors rank for the same term. If two or three competitors all rank for a keyword and you do not, that is a much stronger signal than one competitor ranking for it.

Either way, still run the export through Claude for the clustering step. The tools tell you where the gaps are. The AI is what turns 340 rows into five categories you can actually assign to someone.

One more field worth pulling if your tool supports it: whether a keyword triggers an AI Overview and whether a competitor gets cited in it. That is your AI visibility gap, and it often points to competitors you would not have expected.

Question to Answer:

Do you have a paid tool already, or should you run the free Keyword Planner and SpyFu version first?

9Score the clusters and decide what is worth doing

Not every gap deserves your time. Score each cluster across four factors and total the points.

Anything scoring 8 or higher goes to the top of the list. In plain terms you are looking for strong intent, low difficulty, and weak competing pages.

Do not get pulled in by search volume alone. If a keyword does not line up with what you actually sell, cut it, no matter how big the number looks. A high-volume term that brings you visitors who will never buy is worse than no traffic, because it costs you the time you could have spent on something that converts.

Question to Answer:

Which three clusters score highest, and does each one connect to something you actually sell?

10Build, update, merge, or skip

Every scored cluster ends in one of four decisions. Check the search results before you pick a page type, because if the top results are all step-by-step guides, a service page is not going to rank there no matter how good it is.

Then match the gap type to the action:

The rule that saves the most wasted work: check Google Search Console before you publish anything new. If you already get impressions for a keyword, even at position 40, updating that page beats publishing a second one every time. You already have something Google considers relevant. Build on it.

And never publish two pages for the same intent. You are not doubling your chances, you are splitting them.

Question to Answer:

For your top cluster, does Search Console already show impressions on an existing page you could update instead?