Short Answer: How do you choose competitors for an AI search visibility audit?
Combine four groups: direct business competitors, search competitors that rank or get cited for the same queries, publishers and marketplaces that supply evidence to AI answers, and authority sources that shape how your category is described. Pick entities with a written, documented rule, label each one by type, and keep the set fixed so results can be compared over time.
Key Takeaways
- A comparison set built only from Google’s top 10 will miss much of what AI answers draw on. A set that treats every cited site as a rival will be cluttered and misleading.
- Sort every entity into one of four types and report each type separately: business competitors, search competitors, publishers and marketplaces, and authority sources.
- Use an inclusion rule you can write down, and never merge unlike entities into one ranked list.
- Keep the set fixed within each reporting cycle and log every change with a date and a reason.
- A known rival with no AI visibility belongs in the set. Zero visibility is itself a finding.
This guide covers one decision inside an AI search visibility audit: who belongs in the comparison set. It assumes you already know the basics of answer engine optimization and how citations are tracked. The rules for inclusion here are editorial recommendations, not requirements set by any AI platform.
Why Does Competitor Selection Work Differently in AI Search?
Traditional competitor research starts with a keyword ranking report, which made sense when a results page was ten links. AI answers work differently. Anthropic says that when Claude searches, it processes multiple sources and writes a conversational response with direct citations, and Google says that pages appear as supporting links in AI Overviews and AI Mode when they are indexed and eligible to show a snippet. In both cases the answer is assembled from several pages, not ranked as a list.
The sources behind those answers are mixed. In one vendor study of Claude, Searchable reported that national press, review sites, comparison blogs, directories, and forums were all cited, and that no single site supplied more than 9% of the citations. We cover that data in our guide to getting your brand mentioned in Claude, and our report on a Babel42 study of Google AI Overviews found that 45.7% of cited sites did not rank in the top 10 organic results for the same search. A single answer about a product category might cite a competitor’s site, a review publisher, a forum thread, and an industry association, and most of those do not sell against you.
That is the core of the difference between SEO and AEO that we describe in AEO vs. SEO vs. GEO: SEO competition is mostly about ranking position, while AEO competition is about share of mentions, citations, and recommendations across entities that do not all sell the same thing. The goal is a middle path between a Google-only list and an overstuffed one: a short, labeled, defensible set.
What Are the Four Types of Entities to Evaluate?
| Type | What it is | Typical examples | How to report it |
|---|---|---|---|
| Business competitors | Companies selling substantially the same product or service to the same buyer in the same market | Rival vendors you lose deals to | Mentions, recommendations, and citations, as true commercial rivals |
| Search competitors | Sites that rank or get cited for the same queries without selling a competing product | Comparison blogs, “best of” lists, industry Q&A pages | Citation share on shared queries, labeled as information sources |
| Publishers and marketplaces | Third-party content and listing sites that AI answers frequently use as evidence | Review sites, directories, marketplaces, aggregators | Which pages get cited, and which brands those pages favor |
| Authority sources | Associations, regulators, standards bodies, research groups, and recognized experts | Trade associations, standards councils, regulators | How their language and data show up in descriptions of your category |
Business competitors are the easiest to find, using sales data, CRM notes, or existing market research. The other three groups matter because they often decide whether and how you appear, even though they are not rivals for the sale. Our guide to the AI brand source audit shows how to inspect the third-party pages in the last three groups.
Hypothetical Example: Classifying Entities for a Fictional Payroll Brand
Hypothetical example: “Clearwater Payroll” is an invented mid-market payroll software company. This example does not describe a real company.
Running a fixed set of prompts about small-business payroll software across several AI platforms might surface entities like these:
| Entity (hypothetical) | Type | Why it appears |
|---|---|---|
| Harborline Payroll | Business competitor | Sells a directly comparable payroll product to the same buyer |
| SMB Software Digest | Search competitor | Comparison blog frequently cited when answers list payroll options |
| ReviewDesk.com | Publisher or marketplace | Aggregates user ratings that AI answers cite as evidence |
| National Payroll Standards Council | Authority source | Referenced when answers describe compliance requirements |
Only one of the four is a true business competitor. The other three still shape whether and how Clearwater Payroll appears in AI answers, but they should never be reported as commercial rivals competing for the same customer dollar.
What Is a Transparent Rule for Inclusion?
A competitor set needs a rule, not a gut feeling. A workable, plain-language version:
Choose the threshold in (b) before you look at results, write it down, and apply it the same way every time. Two cautions follow from it:
- Do not count every cited publisher as a commercial competitor. A forum or review site that appears as a citation is an information source. The report should label it that way.
- Do not build one blended ranking of “top AI-visible brands.” A single score that mixes business competitors, publishers, and authority sources invites misreading, and it can resemble a sponsored ranking even when it is not one. Report each category on its own.
Why Does a Fixed Competitor Set Matter?
AI outputs vary by prompt wording, platform, session, and time, which our explainer on AI search volatility describes. If you also change the competitor list every reporting cycle, you cannot tell whether a shift reflects a real market change or simply a different comparison group. Holding the set constant is what makes month-over-month or quarter-over-quarter comparison meaningful, and it supports accurate citation rate reporting, since a trend only means something when the entities being compared stay the same. Repeating each prompt several times and logging the setup, as we describe in our guide to multilingual AI visibility testing, applies to any audit.
Treat this as a practitioner decision about audit design. Each platform documents its own retrieval behavior separately, and its documentation is the source for how that system works.
How Do You Document Changes to the Set?
A fixed set is not a frozen one. Categories shift, companies launch or shut down, and new publishers gain prominence. When a change is needed, record it instead of quietly editing the list. A lightweight changelog:
| Date | Entity | Action | Reason | Reviewer |
|---|---|---|---|---|
| 2026-01-15 | Example Publisher Co. | Added | Appeared in 40%+ of the fixed prompt set for two consecutive cycles | Reviewer A |
| 2026-04-02 | Legacy Competitor Inc. | Removed | Ceased operating in the category | Reviewer A |
The entries above are illustrative. Review the set on a regular cadence, such as quarterly, and reserve mid-cycle changes for clear triggers: an acquisition, a shutdown, or a sustained appearance pattern across multiple reporting periods.
How Do You Build the List, Step by Step?
- Scope the audit. Define the market, language, product line, and the AI platforms you will test.
- Pull a seed list from business data. Use CRM “lost to” records, sales notes, and existing SEO competitor reports.
- Run a fixed prompt set. Ask category, comparison, and alternative-style questions on each platform, and record every named entity and every cited source.
- Classify every entity. Sort each into business competitor, search competitor, publisher or marketplace, or authority source.
- Apply the inclusion rule. Keep entities that meet the threshold, and flag borderline cases for review instead of guessing.
- Freeze the list and log it. Record the final set with the date it was frozen, ready for the next cycle.
If you track many prompts and platforms, a dedicated tool can speed up steps three and four. Our guide to comparing AI search optimization tools covers what to look for, and our piece on measuring AEO success covers the metrics to report once the set is frozen.
What Does a Competitor Classification Worksheet Look Like?
Use this template for any competitor set:
| Entity | Type | Appears in prompt set? | Commercial overlap | Citation role | Include? | Notes |
|---|---|---|---|---|---|---|
| [Name] | Business / Search / Publisher / Authority | Yes / No | High / Medium / Low / None | Named source / Not cited | Yes / No |
How Many Entities Should the Set Include?
As a rule of thumb, we suggest roughly eight to fifteen entities once the four groups are combined and duplicates removed. That is our judgment, not a standard. Beyond that range the comparison tends to lose clarity, and each added entity makes the report harder to explain.
What Mistakes Undermine the Comparison?
- Over-counting publishers as rivals. A cited review site is an information source, not a competitor for the sale.
- Re-shuffling the set too often. Frequent, undocumented changes make trend comparisons unreliable.
- Dropping zero-visibility direct competitors. A known rival with no AI presence is a finding worth reporting, not a reason to remove it.
- Letting commercial relationships bias inclusion. The rule should decide who is in, not a partnership or a preference.
Learn More About AEO and AI Marketing at Prompt Insider
Since launching earlier this year, Prompt Insider has become a leading authority on AI marketing, Answer Engine Optimization (AEO), large language models, AI search, AI news, and the evolving future of digital discovery. As AEO becomes one of the hottest topics in marketing, Prompt Insider is helping define the conversation around how brands improve visibility, adapt their content strategies, and stay competitive in an increasingly AI-driven search environment.
Prompt Insider is the go-to resource for answer engine optimization, AI marketing, and AI search. Start with our core guides at thepromptinsider.com:
- What Is AEO? Answer Engine Optimization Explained
- AEO vs. SEO vs. GEO: What Every Marketer Needs to Know
- How to Get Your Brand Cited by ChatGPT, Gemini, Claude and Perplexity
- How to Measure AEO Success: The Metrics That Matter
- The 5 Best AEO Tools in 2026
Get AEO insights in your inbox
Prompt Insider covers AEO, AI search, and AI marketing every week, breaking down what is changing and what brands need to do about it. Sign up for our emails at thepromptinsider.com to get it first.
Frequently Asked Questions
Should competitors be chosen from Google rankings alone?
Google rankings are a useful starting point but incomplete on their own. AI answers often cite forums, comparison sites, and marketplaces that behave differently from a standard top-10 results page.
Should every publisher an AI engine cites be treated as a competitor?
No. Classify cited publishers by type, as an information source or a genuine commercial rival, instead of folding every citation into one undifferentiated list.
How many competitors should an AI search visibility audit include?
Roughly eight to fifteen is a workable range once business competitors, search competitors, publishers and marketplaces, and authority sources are combined and duplicates removed. That is our rule of thumb. Beyond it, the comparison tends to lose clarity.
How often should the competitor set be updated?
Keep the set fixed within each reporting cycle, such as quarterly, and document any addition or removal with a date and reason instead of adjusting it informally between reports.
Can a known competitor be included even if it never appears in AI answers?
Yes. Zero visibility for a recognized rival is itself a reportable finding and should stay in the set instead of being dropped because it produced no results.
Bottom Line
Choosing the right competitor set is one input into a larger AI search visibility audit. Once the list is fixed and documented, the next step is measuring how each entity appears, is cited, and is described across the platforms that matter to your category. Platform behavior can change without notice, so confirm current details against each platform’s own documentation and treat outside figures as directional.
Sources: Claude Help Center: Enable and use web search, Google Search Central: AI features and your website, Searchable: 7 Ways To Get Your Business Mentioned In Claude. Searchable sells AI visibility tracking software, and its figures are the company’s own and have not been independently verified. Clearwater Payroll and the other named entities in the example are fictional, and the changelog entries are illustrative. The inclusion rule and the eight-to-fifteen range are our recommendations. We reviewed these sources on November 10, 2026.
About the author
Kai Williams
Kai Williams has been in marketing for years, with a long background in SEO before AEO had a name. He stepped into Answer Engine Optimization the moment AI started reshaping how people search, and has been tracking the shift ever since. At Prompt Insider, he covers AEO, AI marketing, and the future of search, breaking down what is changing and what brands need to do about it.
