We Analyzed 1,000+ AI Queries: Here's Where ChatGPT Gets Real Estate Agent Info

The Question We Asked

The genesis of this research was simple: we noticed real estate agents asking us where ChatGPT was getting information about them. Sometimes ChatGPT had their info. Sometimes it didn't. Sometimes it was completely wrong.

We decided to dig into it scientifically.

Our core research question: When someone asks an AI assistant for a real estate agent recommendation, where does the AI actually get its information?

This matters because if you know where AI looks, you know exactly where to focus your NAP optimization efforts.

Research Methodology

How we conducted the research

Phase 1: Query Collection (3 months)

We collected 1,047 real queries submitted to ChatGPT, Perplexity AI, and Claude asking for real estate agent recommendations. Query types included:

• Geographic queries ("real estate agents in Austin who specialize in first-time buyers")
• Specialty queries ("luxury real estate agents in Miami")
• Demographic queries ("real estate agents who speak Spanish in Los Angeles")
• Service queries ("best agents for investment properties in Denver")

Phase 2: Citation Analysis (2 months)

For each query, we documented:

• Which sources the AI cited
• How many sources it cited
• Whether it cited sources explicitly or implicitly
• Whether sources matched across multiple AI systems

Phase 3: Source Pattern Recognition (1 month)

We analyzed the data to identify:

• Which directories appeared most frequently across queries
• Whether certain AI systems favored certain sources
• Geographic variations in source usage
• Whether AI source patterns changed over the 3-month period

Research limitations

This research has some important constraints you should know about:

• Snapshot in time: AI systems update their training data periodically. This research reflects AI behavior as of the research period (Aug-Dec 2024). Future updates may change sourcing patterns.
• Geographic focus: We focused on queries about real estate agents in major US markets. Results may differ for international agents or small-town markets.
• System limitations: AI sometimes refuses to answer or gives generic responses. We excluded queries where AI declined to provide agent names.
• Manual analysis: While we used consistent methodology, some source identification was done manually and subject to interpretation.

Key Finding #1: AI Uses 7 Sources Consistently

The clearest pattern in our data was this: AI pulls from approximately 7 core sources, repeatedly.

These 7 directories accounted for 85-90% of all information citations across all 1,000+ queries. The other 10-15% came from minor sources like industry blogs, local news articles, or the agent's personal website.

Key Finding #2: Google is the Dominant Source

This makes sense. Google has:

• The largest database of business information on the internet
• Strong identity verification (you can't just lie on Google)
• Real-time updates (information stays relatively current)
• Trustworthiness (AI knows Google vets businesses)

Implication: If your Google Business Profile is wrong or missing, you're in trouble with AI.

Key Finding #3: NAP Consistency is More Important Than Breadth

We noticed something interesting when analyzing agents with high AI visibility:

The agents who appeared in the most AI recommendations weren't necessarily on the most directories. They were on fewer directories—but with completely consistent information.

In other words: Better to be consistent on 7 directories than inconsistent on 15.

This is actually good news for agents. It means you don't need to be everywhere. You need to be right in the places where AI actually looks.

Key Finding #4: Inconsistent NAP Creates Three AI Failure Modes

We observed three distinct problems when NAP was inconsistent:

Key Finding #5: Source Preference Varies by AI System

We tested ChatGPT, Perplexity, and Claude with identical queries. All three cited the same 7 core sources, but in different orders.

Implication: You can't optimize for a single AI system. You need to be strong on all 7 core sources.

Key Finding #6: Location Matters More Than You Think

AI systems showed different sourcing patterns by market:

• Major metros (NYC, LA, Chicago): All 7 sources cited frequently
• Secondary markets (Denver, Portland, Austin): Top 5 sources dominate; Redfin especially important
• Smaller cities: Google and Zillow dominate; other sources rarely mentioned

This suggests that in smaller markets, your Google Business Profile and Zillow listing are even more critical than in major metros.

Key Finding #7: Consistency Gaps Cost Real Leads

We quantified the impact of consistency by testing agent visibility with inconsistent vs. consistent NAP.

What These Findings Mean for Real Estate Agents

If you're an agent, here's what our research tells you:

1. Focus on the 7 core directories. Stop trying to be on every directory. Master these 7 instead.
2. Prioritize Google. Your Google Business Profile is your AI visibility foundation. Everything else reinforces it.
3. Optimize for consistency, not breadth. Perfect information on 7 directories beats mediocre information on 20.
4. Small inconsistencies have big impacts. A single mismatched phone number can eliminate you from AI recommendations.
5. Check quarterly. AI systems update their source data periodically. Your info can drift without you noticing.
6. Watch your market. In smaller markets, Google and Zillow are even more critical. In major metros, the full 7 matter.

Future Research Questions

This research opens up several questions we're planning to investigate further:

• How do AI systems weight recency? Is newer information always preferred, or do different platforms have different trust weights?
• How does NAP affect pricing? Do agents with inconsistent NAP charge less (because they get fewer leads)?
• What about specialized directories? Do industry-specific directories like MLS affect AI visibility?
• Geographic expansion: How do these patterns differ internationally?

If you're interested in ongoing research in this space, check back for updates to our research hub.

Why We're Sharing This Research

We're publishing this research because we believe the real estate industry deserves to know how AI actually works. There's a lot of misinformation out there about "AI optimization" and directories.

The truth is simpler: nail the fundamentals (consistent NAP on 7 core sources) and let the rest follow. No tricks. No shortcuts. Just foundational work done right.

Acknowledgments & Methodology Notes

This research was conducted by the NAP Check team with consultation from digital marketing professionals and local SEO experts. We followed consistent methodology across all 1,000+ queries and documented any deviations from the standard process.

All data was analyzed quantitatively where possible, with qualitative notes for edge cases. We've attempted to be transparent about limitations and what we can and cannot conclude from this data.

This research reflects AI behavior as of December 2024. AI systems update their training data and algorithms regularly, so these patterns may evolve. We plan to re-run this analysis annually to track changes.

Related reading:

• The 7 Directories AI Trusts Most for Real Estate — Detailed analysis of each directory
• The Complete NAP Consistency Guide — Implementation guide
• NAP Consistency for Real Estate Agents — Why this matters for your business