Turn competitor Google keywords into AI prompts

The blank prompt list is where many AI visibility projects stall. You sign up, see an empty monitor and a cursor, and have to decide which questions buyers ask ChatGPT. Most people guess, write twenty prompts about their own brand and learn very little.

There is a better starting point you already own: keywords. A domain that ranks on Google for a term has proven there is demand and content that matches it. Those terms are a ready-made source for prompts, including for a competitor's domain.

Why keywords are a good seed

Keyword data has three advantages over guesses.

It reflects real demand. Someone typed these phrases. You are not inventing a question nobody asks.

It covers the topics you actually compete on. The keywords a competitor ranks for describe the territory they have claimed.

It is specific. "Best invoicing software for freelancers" is a prompt. "Invoicing" is not.

The caveat is that keywords are not prompts. A keyword is often two to four words. A prompt is a question or request. Converting one into the other is the work.

The manual method

Here is a process you can run with any rank tracker or Search Console:

Collect keywords. Export your own ranking keywords, and do the same for two or three competitors if your tool allows it.
Filter for commercial intent. Keep terms with comparison, best, alternative, price, review or how-to signals. Drop navigational brand queries.
Rewrite as questions. "crm for small business" becomes "What is the best CRM for a small business with five people?" Add a clause of context, because AI prompts tend to be longer and more specific than keywords.
Group by topic. Ten prompts on invoicing and ten on payroll are easier to read than a flat list.
Add intent labels. Mark which prompts are category searches, brand searches and competitor comparisons.
Test a sample. Run 10 prompts and see whether answers name brands. If an engine answers with no brands, deprioritise that prompt.

Keep the list small at first, around 30 to 50 prompts. A tight set you review weekly beats a large set you never read. Our piece on prompt volume versus keyword search volume explains why the two numbers do not map one to one, so do not expect keyword rank to predict AI visibility.

What to look for in a tool

If you automate this, look for three things:

Keyword lookup for any domain, not only your own, so a competitor's rankings are available.
Conversion into prompt suggestions, not just a keyword export.
A direct path from suggestion to monitor, so you do not copy and paste between sheets.

Our guide on tools that track keyword rankings in AI search results compares products that connect the two worlds.

Where Promptwatch fits

Promptwatch built this into Prompt Explorer. Per the September 19, 2026 changelog entry, Prompt Explorer has a new Ranking keywords tab. You enter a page or domain, and Promptwatch looks up the Google keywords it ranks for, then turns them into prompt suggestions you can add to a monitor.

The tab sits next to the other sources: Search Console, People also asked, seed query and website page. That gives you several ways in. Use Search Console for what your own site already earns impressions for, People also asked for question-shaped phrasing, and Ranking keywords when you want a competitor's territory.

Two related changes help new projects. On September 21 the changelog added starting prompts when you create a monitor, so a new monitor starts collecting answers right away instead of sitting empty. New projects also get two extra monitors after onboarding, each aimed at a different country or city.

The changelog does not state a plan for Ranking keywords, so I will not attach one. For context, Prompt Volumes show search volumes, difficulty scores and citation rates across AI search, Google and Bing, so after you add keyword-derived prompts you can sort them by demand rather than by the order you typed them. Personas and topics help you organise the result, and you can find a walkthrough in our post on turning Search Console topics into prompts.

Why choose it for this job? Because the step that usually breaks is the conversion from "keyword list" to "monitor with prompts". Having the lookup, the rewrite into suggestions and the monitor in one product removes the spreadsheet in the middle.

How to use a competitor's keywords without copying them

A competitor's keyword list is a map, not a plan. Use it with care.

Look for gaps first. Compare their keywords to yours. Terms they rank for and you do not are candidate topics. Terms you both rank for show where answers will be contested.

Prioritise by intent. A comparison or "best" prompt is worth more tracking effort than an informational one, because the answer names brands.

Check who gets named. Run the new prompts and see whether the competitor is named, cited or absent. Ranking on Google does not guarantee a mention in AI answers, and the gap itself is useful: a page that ranks well but is never cited tells you something is off. Our post on pages that rank on page one but are not cited in ChatGPT takes that further.

Do not chase everything. Pick one topic cluster, track it for a few weeks, and widen only if you learn something.

A starting checklist

Pick one competitor domain and your own.
Pull ranking keywords for both.
Keep commercial and comparison terms.
Convert 30 to 50 into questions with a clause of context.
Add them to a monitor, tag by topic and intent.
Review after two weeks and drop prompts that produce no brand names.

The goal is a prompt list grounded in demand rather than opinion. Start from what already ranks, rewrite it as the question a buyer would ask, and let the answers tell you which ones matter.