# Where do Genezio, Peec AI and Profound get your competitor list, and why does it change your numbers?

URL: https://theansweroptimizer.com/genezio-peec-ai-profound-competitor-list/
Published: 2026-10-20

> Share of voice, rank and win rate depend on which rivals a tool counts. We compare where Genezio, Peec AI and Profound get their competitor lists, how they merge name variants, and which list each metric divides by.

Generative Engine Optimization (GEO) tools measure how often AI assistants such as ChatGPT mention your brand. Most also compare you with rivals. That comparison depends on a question few dashboards answer: which rivals does the tool count?

This article compares [Genezio](https://genezio.com), [Peec AI](https://peec.ai) and [Profound](https://www.tryprofound.com) on that one point. It follows the competitor list from where it starts, to how name variants are joined, to which list each metric uses.

In-product observations come from our review of Peec AI, Genezio and Profound, in one signed-in account per product.

## Why the competitor list matters as much as the metric

Share of voice, average position, win rate and rank lists have one thing in common. Each is calculated against a set of brands. Change the set and the number changes, even if the AI answers stay the same.

Each of the three tools works with two sets, not one:

- **The field** is every brand the tool detects in an AI answer.
- **The tracked set** is the brands you see and benchmark on the dashboard.

Some metrics use the field and others use the tracked set. This can differ between metrics inside the same tool.

**What this means:** two tools can both report "you rank #2" while counting different rivals. Before comparing numbers, you need to know where each list came from.

## Where does each tool's competitor list come from?

**Profound** sets the tracked list at onboarding: "your Profound team configures your primary competitive set". Its visibility score, however, counts any brand named in an answer, "not only your configured competitors", unless you exclude that brand. Users can exclude or include brands afterwards. In our review, the starting set looked prefilled at setup rather than built from brands observed in answers.

**Peec AI** suggests competitors for you to accept or reject. In our review, the first set of suggestions did not come from AI answers. It was prefilled behind the scenes during onboarding. That is the step where you connect your domain and Peec AI builds a project profile of your business. Our inference is that a language model, the kind of AI behind ChatGPT, is asked who your competitors are. Later, once your prompts run, Peec AI also suggests brands mentioned alongside yours "when they appear at least twice", with a mention count next to each.

**Genezio** automatically finds competitor brands when it analyzes conversations. You can then promote a brand to "a primary competitor" or demote it so the analysis "gives them less importance". Genezio does not document a detection threshold, or whether a demoted brand still counts in share of voice.

So none of the three relies on a fixed list alone. They differ in where the tracked set starts and who curates it.

## What happens when one competitor appears under two names?

Once a brand is on the list, is it counted as one brand? AI answers often name the same company in several ways. If the variants are not joined, one company's mentions are split across two entries. Detection itself is covered in [how GEO tools decide an answer mentions your brand](https://theansweroptimizer.com/geo-tool-brand-mention-detection/).

**Profound** documents a Name Matching feature with suggestions and a merge action. It combines "string matching, fuzzy matching, and an LLM review step". An LLM is a large language model, the kind of AI behind ChatGPT. Profound warns that without matching, "each variation counts as a separate entity". It also says it "can't always tell" a parent brand from a sub-brand.

**Peec AI** gives each brand a tracked name, aliases and an optional regular expression, which is a text pattern for matching names. Saving changes triggers a recalculation across existing chats. In the account we reviewed on 14 October 2026, Peec AI showed duplicate competitor entries under name variants, such as "Scrunch" and "Scrunch AI". Each pair had identical figures. That is one account on one day, not a defect rate. Peec AI documents no merge of two entries. The closest documented route is to add the variant as an alias on one entry and delete the other.

**Genezio** has a merge for competitor entries, so all mentions are under one brand. It also lets you attach sub-brands or acquired companies to your own brand, so they are not counted as rivals.

## Which list does each number divide by?

With a clean list, the remaining question is which list each metric uses. Here the three tools differ most.

**Peec AI**'s rank metrics use the whole field. Its position metric "counts every brand detected in a response, not only your tracked competitors". Win rate is the share of answers where you are ranked first. If an untracked brand ranks above you, "the chat will not count as a win".

Peec AI's share of voice is less clear. Its documentation describes the comparison group in two ways: "all brands mentioned" in one place, and tracked competitors or tracked brands in others. In practice, you cannot tell from the docs which rivals divide your share, so it is a question to put to Peec AI.

**Profound**'s metrics use the field minus exclusions. Visibility and share of voice are calculated per AI model, then averaged with equal weight. Average position is averaged only over the models where your brand appeared. Excluded brands are removed before the calculation.

Profound's documentation and our review differ on its rank list. The docs say the Visibility tab ranks you "relative to competitors", and that a filter chooses "which of the brands you track appear in its charts and rank lists". In our review on 14 October 2026, the rank we saw covered the whole detected field, not only tracked brands. Check which brands your own rank list includes.

**Genezio**'s share of voice divides your mentions by mentions of "your brand and the competitors". Those competitors are the ones it has detected or you have promoted. The documentation notes the metric is zero-sum, and suggests checking whether a new brand appeared if your share falls. **Our inference:** because competitors are added automatically, a newly detected brand can lower your share without any change on your side. Genezio's own product description says it ranks you against tracked competitors only, not the whole field.

Ranking first is also not the same as being recommended, a difference covered in [does ranking first mean the AI recommends you?](https://theansweroptimizer.com/geo-tool-rank-first-vs-recommended/)

## Do brand-perception comparisons use the same competitors?

Perception features describe how AI characterises your brand, and they can bring in yet another list.

**Peec AI**'s Market view runs on its own set of AI answers, not your tracked prompts. Its competitors come from the industry you select, which "decides which brands you are measured against". Changing the industry later does not rerun the analysis automatically, and re-runs are limited. Peec AI's changelog entry of 13 October 2026 adds one-click tracking of untracked brands from Brand Perception.

**Profound** reports sentiment from prompts you assign the Sentiment type. Its documentation says it "analyzes the response for any brands mentioned", and that "Sentiment prompts run across your full competitor set". The docs do not say whether brands outside that competitor set appear in the sentiment views.

**Genezio** offers a SWOT drawer, short for strengths, weaknesses, opportunities and threats. The docs say it opens "where a competitor appears" and "compares your brand directly with that competitor". Separately, its Comparer agent tests how AI answers compare your brand with competitors you select.

**What this means:** a perception report may not use the same brand set as the visibility dashboard. More on perception question sets is in [which questions brand perception is based on](https://theansweroptimizer.com/geo-tool-brand-perception-which-questions/).

## What should you check before comparing yourself with rivals?

The differences above come down to a few checks. None of the three publishes an error rate for competitor or brand detection, so the checks fall to you.

*Each tool builds and applies its competitor list differently, so the same metric name can rest on different rivals. Cells are from vendor documentation read in October 2026 unless marked otherwise.*

| Question | Genezio | Peec AI | Profound |
|---|---|---|---|
| Where the tracked set starts | Added automatically from answers and cited sources | Prefilled at onboarding from the project profile, not from answers (our review); later, suggestions from brands seen in answers twice or more | Configured by Profound's team at onboarding; users can exclude or include brands |
| Merging name variants | Merge for competitors; no own-brand alias field found | Aliases and pattern per brand; no merge of two entries | Name Matching with suggestions and merge |
| What rank is measured against | Tracked competitors only (vendor's own description) | Every detected brand | Metric denominator: every detected brand minus exclusions (documented). Rank list filtered to tracked brands (documented). Whole-field rank observed in our review, 14 Oct 2026 |
| Effect of list edits on past numbers | Not documented | Recalculation runs; effect on past share of voice not spelled out | Excluding a brand removes its past mentions from every report; switching to an include list does the same (documented). Other edits not stated |

Five questions work with any of the three:

1. **Where did each name come from?** Ask whether it was observed in answers or added at setup.
2. **Are variants merged?** Check for duplicate and unmatched names in the first weeks.
3. **Which list does each metric use?** Ask the vendor for the denominator of share of voice and rank.
4. **Do list edits rewrite history?** Profound says excluding a brand "removes its past mentions from every report". Export your numbers at reporting time, as covered in [whether past numbers can change](https://theansweroptimizer.com/geo-tool-past-numbers-change/).
5. **Can you see the untracked field?** If you can [export raw answers](https://theansweroptimizer.com/geo-tool-raw-answer-export/), you can check which brands were left out.

In short, each tool has a documented strength and a documented gap. Profound documents how exclusions rewrite every past date, but its team sets the starting list. Peec AI ranks you against every detected brand, but its share-of-voice wording is inconsistent. Genezio grows the list from answers, but does not document what demotion or list edits do to your numbers.

None of these choices makes a tool more accurate on its own. They decide which rivals a number counts. For broader fit, see the pairwise comparisons of [Genezio and Peec AI](https://theansweroptimizer.com/genezio-vs-peec-ai-workflow-fit/), [Genezio and Profound](https://theansweroptimizer.com/genezio-vs-profound-measurement-enterprise/), and [Peec AI and Profound](https://theansweroptimizer.com/peec-ai-vs-profound-monitoring-reporting-trade-offs/).