How to Track Your Brand Visibility in ChatGPT

The Problem

883 million people use ChatGPT monthly. Many of them are your potential customers.

When they ask "best [your category]" or "which [product type] should I choose", does ChatGPT recommend you?

You probably don't know. Unlike Google Search Console, ChatGPT provides zero analytics.

This guide shows you how to track your brand visibility in ChatGPT — manually (free) or automated (paid).

Why ChatGPT Visibility Matters

The Data

• 60.7% market share of AI search
• 700M+ weekly queries
• 47% of B2B buyers name ChatGPT as their preferred LLM for research
• 4-5x higher conversion for AI-referred traffic vs traditional search

The Blind Spot

When someone asks ChatGPT "best marketing agency for B2B SaaS", ChatGPT recommends 2-3 agencies.

If you're not one of them, you don't exist.

But you have no way to know:

• Which prompts prospects are using
• Whether ChatGPT recommends you
• Which competitors ChatGPT prefers
• Why ChatGPT chose them over you

This is the new marketing blind spot. And it's bigger than most brands realise.

Why Retrieval ≠ Citation

A 2026 AirOps study analysed 548,534 pages retrieved across 15,000 ChatGPT prompts and found that 85% of pages ChatGPT discovers never appear in the final answer.

ChatGPT works in three stages:

• Discovery: ChatGPT identifies potentially relevant pages — often from Google's index. Ranking #1 on Google makes you 3.5× more likely to be cited than pages outside the top 20.
• Expansion: 89.6% of prompts trigger two or more follow-up searches. ChatGPT actively researches the question before answering. Most brands get dropped here.
• Selection: Only ~15% of retrieved pages make the final answer. The pages that survive: 44.2% of citations come from content in the first 30% of the page. 68.7% of cited pages have a clear H1→H2→H3 heading structure.

This means tracking whether ChatGPT mentioned you isn't enough. You need to track whether it cited you — and understand whether you're getting dropped at the expansion or selection stage.

Manual Tracking Method (Free)

Step 1: Define Your Core Prompts

Brainstorm 10-20 prompts your buyers might use:

Category-level:

• "Best [your category]"
• "Top [your category] for [use case]"
• "[Your category] for [industry]"
• "[Your category] in [location]" (for local businesses)

Comparison:

• "[Your brand] vs [competitor]"
• "Is [your brand] better than [competitor]"
• "Alternatives to [competitor]"

Decision-stage:

• "Which [your category] should I choose"
• "How to choose a [your category]"
• "[Your category] buyer's guide"

Step 2: Query ChatGPT

Open ChatGPT and enter each prompt exactly as written.

Important: Use a fresh session for each prompt (or clear context with "Forget everything we discussed").

Why? ChatGPT's responses are contextual. If you ask "Tell me about Decyde" followed by "Best AI visibility tool", the second answer is biased by the first.

Step 3: Record Results

For each prompt, record:

1. Mentioned? Yes/No
2. Position: 1st recommendation, 2nd, 3rd, or honorable mention?
3. Sentiment: How is your brand described?
   • Positive: "Industry leader", "Best for X"
   • Neutral: "Offers X service"
   • Negative: "Newer player", "Limited features"
4. Competitors mentioned: Who else was recommended?
5. Reason given: Why did ChatGPT recommend you (or not)?

Step 4: Calculate Your Mention Rate

Mention Rate = (Prompts where you appear) / (Total prompts) × 100

Benchmarks:

• 0-20%: Invisibility problem
• 20-40%: Average
• 40-60%: Competitive
• 60%+: Category leader

Step 5: Repeat Weekly

Track the same prompts every week to identify trends:

• Did your mention rate improve?
• Did your position change (2nd → 1st)?
• Did sentiment shift (neutral → positive)?
• Did new competitors appear?

Time investment: ~20 minutes per week for 20 prompts.

Automated Tracking Method (Paid)

Manual tracking works for 10-20 prompts. But what if you need to track 50+? Or track daily instead of weekly?

Tools Available

Cost-Benefit Analysis

Manual tracking:

• Time: 20 minutes/week = 1.3 hours/month
• Cost: $130/month (at $100/hour labor rate)
• Limitations: Small prompt set, no historical data, manual entry errors

Paid tool:

• Time: 5 minutes/week (just review the report)
• Cost: $99-499/month
• Benefits: 50-500 prompts, historical trends, competitor benchmarking

Breakeven: If your time is worth $75+/hour, paid tools are cheaper than manual tracking.

What to Do If You're Not Mentioned

If ChatGPT doesn't recommend you, the fix depends on why.

Reason #1: "I'm not familiar with that brand"

Diagnosis: Entity recognition problem. ChatGPT doesn't know you exist.

Fix:

• Get listed in "Best [category] tools" articles
• Earn reviews on G2, Capterra, Trustpilot
• Get press coverage (even local news counts)
• Add structured data (schema markup) to your website
• Build a Wikipedia page or knowledge panel

Timeline: 60-90 days to see improvements.

Reason #2: "For [use case], I'd recommend [competitor] instead"

Diagnosis: ChatGPT knows you exist but prefers competitors for specific use cases.

Fix:

• Create comparison content ("[Your brand] vs [competitor]")
• Publish use case-specific content ("Best [your category] for [use case]")
• Highlight differentiators on your homepage
• Earn citations in niche publications (not just general tech press)

Timeline: 30-60 days to see improvements.

Reason #3: "I don't have access to current information about that brand"

Diagnosis: ChatGPT's training data doesn't include recent info about you.

Fix:

• Use ChatGPT Search mode (available in ChatGPT Plus/Pro)
• Publish fresh content monthly (blog posts, case studies, press releases)
• Update your website with "Last updated: [date]" timestamps
• Get mentioned in recent articles (2025-2026, not 2022-2023)

Timeline: Immediate (if using Search mode), 30-60 days otherwise.

Reason #4: "[Your brand] is a newer/smaller player"

Diagnosis: Perception problem. ChatGPT sees you as less established.

Fix:

• Highlight social proof (number of customers, years in business, funding raised)
• Earn third-party validation (awards, rankings, case studies)
• Publish thought leadership (guest posts, conference talks, webinars)
• Focus on a niche ("Best [category] for [vertical]" vs "Best [category]")

Timeline: 90-180 days (reputation takes time).

Advanced: Prompt Engineering for Better Data

Not all prompts are equal. Here's how to craft prompts that reveal more insights.

1. Use Buyer Language

Weak: "List AI visibility tools"

Strong: "I'm a law firm partner looking to improve our visibility in ChatGPT. Which tools should I use?"

Why: The second prompt reveals how ChatGPT tailors recommendations by buyer persona.

2. Add Constraints

Weak: "Best AEO tools"

Strong: "Best AEO tools for a $99/month budget"

Why: Constraints force ChatGPT to prioritize. You learn which attributes it values.

3. Ask for Comparisons

Weak: "Tell me about Decyde"

Strong: "Compare Decyde vs AIclicks for a law firm"

Why: Comparison prompts reveal your strengths and weaknesses relative to competitors.

4. Request Citations

Weak: "What's the best [category]"

Strong: "What's the best [category] according to recent reviews"

Why: Forces ChatGPT to cite external sources, revealing which publications it trusts.

Case Study: How One Law Firm Tracked (and Improved) ChatGPT Visibility

Background

Mid-sized IP litigation firm in Chicago. Ranked #1 on Google for "IP litigation Chicago" but invisible in ChatGPT recommendations.

Month 1: Baseline Tracking

They manually tested 15 prompts:

• "Best IP litigation firm in Chicago"
• "Top patent attorneys in Illinois"
• "IP litigation firm for startups"
• [12 more variants]

Results:

• Mentioned: 2 out of 15 prompts (13% mention rate)
• Position: Both times ranked 3rd
• Competitors mentioned: 5 other firms consistently ranked higher

Month 2: Content Creation

They created:

• 1 comparison page: "[Firm Name] vs [Top Competitor]"
• 1 FAQ page: "What to Expect in IP Litigation" (20 questions)
• 3 case studies with specific outcomes

Month 3: Third-Party Validation

They:

• Got listed in "Top 10 IP Firms in Illinois" article
• Earned 5 new reviews on Martindale-Hubbell
• Published a guest post in a legal tech publication

Month 4: Re-Test

They re-tested the same 15 prompts:

• Mentioned: 9 out of 15 prompts (60% mention rate)
• Position: 6 times ranked 1st, 3 times ranked 2nd
• Sentiment: Shifted from neutral to positive ("Leading firm for startups")

Result: 13% → 60% mention rate in 90 days.

Attribution: They tracked 2 new clients who specifically mentioned "finding them on ChatGPT."

Common Questions

Q: How often should I track my prompts?
A: Weekly for the first 3 months, then monthly once you establish a baseline.

Q: Should I use ChatGPT Free or Plus for tracking?
A: Free tier is fine for category-level prompts. Use Plus (with Search mode) if you want to check real-time web citations.

Q: Can I track Perplexity and Gemini the same way?
A: Yes, the manual method works for all platforms. Automated tools vary in platform coverage.

Q: What if ChatGPT hallucinates information about my brand?
A: Document it. You can't directly correct ChatGPT, but you can strengthen your entity definition (structured data, Wikipedia, press mentions).

Q: Is it worth paying for automated tracking?
A: If your time is worth $75+/hour and you want to track 50+ prompts, yes. Otherwise, start with manual tracking.