How do you find your competitors in AI search results?

You find your competitors in AI search by running the unbranded questions your buyers ask across the main AI engines, recording every brand named and every domain cited, and counting how often each one appears. The brands with the highest share of answers are your AI search competitors, and they are often not the same companies your sales team or SEO tools list.

Three kinds of AI search competitors

Type	What it is	Example
Direct competitors	Vendors recommended for the same buyer prompts	Rival products in “best X tools” answers
Unexpected competitors	Adjacent tools, open-source projects or platforms the engine treats as alternatives	A broad platform’s add-on module named instead of your product
Source competitors	Publishers, review sites and communities cited for your topics	Review sites, Reddit threads, industry blogs

All three matter. Direct competitors take recommendations. Source competitors decide which brands get recommended.

Step by step

flowchart LR
  A[Unbranded buyer prompts] --> B[Run on each AI engine several times]
  B --> C[Extract brands and cited domains]
  C --> D[Count and rank by share of answers]
  D --> E[Analyze why each competitor wins]
  E --> F[Plan content and outreach]

1. Use unbranded prompts

Start with category, comparison, problem and use-case prompts that do not mention your brand. See how to choose which prompts to track. Branded prompts tell you how engines describe you, not who beats you.

2. Run them on every priority engine

Engines disagree. A brand can lead in ChatGPT and be absent from Perplexity. Run each prompt several times per engine because answers vary between runs.

3. Extract brands and sources

For every answer, record:

• Each brand named and its position in the answer
• Each URL cited, and the domain it belongs to
• How each brand is described (strengths, weaknesses, ideal customer)

4. Rank by share of answers

Count how often each brand appears across all runs. That gives you a share of voice ranking. Do it per engine and per prompt group, since a competitor may dominate one use case and be invisible in another. Our guide on measuring share of voice in AI answers covers the calculation.

5. Analyze why they win

For the top three to five competitors, look at:

• Which sources cite them. Are they on review sites, roundups or communities where you are absent?
• Which of their pages are cited. Comparison pages, docs, pricing, research reports?
• How they are described. The phrases engines use usually trace back to specific pages.
• Where they are weak. Prompts or engines where they do not appear are openings for you.

A worksheet you can use

Competitor	Share of answers	Top cited sources	Where they are absent
Competitor A	Record here	Review site, own comparison page	Perplexity, implementation prompts
Competitor B	Record here	Reddit threads, analyst list	Pricing prompts
Publisher C	Record here	Own guides cited as sources	Not a vendor, target for coverage

Fill this monthly and keep history so you can see momentum.

Doing it manually vs with tools

Manual research works for a first look: 20 prompts, two engines, a spreadsheet. It does not scale to hundreds of prompts across many engines run repeatedly. AI visibility platforms automate the runs, extract brands and citations, and show trends. GrackerAI, for example, reports per-engine share of voice, position and the cited sources behind each competitor’s mentions. The free AI visibility checker gives a quick snapshot of which brands engines associate with your category.

Turning findings into action

• Competitor cited on a review site or roundup you are missing from: pursue a listing or inclusion
• Competitor’s comparison page cited: publish a fair, specific comparison of your own
• Competitor described with outdated facts: make sure your own facts are clear and consistent everywhere
• Unexpected competitor appears: clarify your category positioning so engines understand the difference
• Publisher dominates your topic: pitch data, quotes or guest content to that publisher

For a deeper method, read AI search competitive analysis.

Questions to answer in your monthly review

Use the same questions every month so the review stays short and comparable:

1. Which three brands appear most often across our prompt set, and has the order changed?
2. Did any new brand appear in more than a handful of answers?
3. On which engine is the gap between us and the leader largest?
4. Which third-party domains are cited most, and are we present on them?
5. Which competitor pages are cited repeatedly, and what format are they?
6. How do engines describe the leading competitor compared with us?
7. Which prompts did we gain or lose since last month, and what changed?

Write the answers in a short memo for product marketing and sales. AI answers often reveal how the market positions each vendor, which is useful well beyond the marketing team.

Common mistakes

• Tracking only the competitors on your sales battlecards
• Relying on one engine or one run
• Ignoring source domains, which explain most of the result
• Treating a single month as a trend

Competitor discovery is the most useful first project in AEO because it shows exactly who you are up against and which sources you need to reach.