Loading…
Content is loading, please wait.The AI recommendation field guide
A name is a starting point.
The evidence is what moves you forward.
Turn a recommendation from ChatGPT, Gemini, Perplexity or Google AI Overviews into a more informed interview. Start with the person behind the name and the sources behind the claims.
Illustrative workflow · not a live chat
AI suggested a real-estate agent.
What should I check next?
When search is used, follow the cited page. Does it support that exact claim about that exact person?
ChatGPT search guidanceOpen Sources when available. Check the supporting passage; a related link alone does not verify a recommendation.
Gemini source guidanceFollow the numbered citation beside a claim. Keep an individual agent’s record separate from a team’s results.
Perplexity citation guidanceRead the supporting links in an AI Overview. Ordinary results and ads are separate placements; an overview may not appear.
Google AI Overview guidanceExplore their source records. These are not AI endorsements.
The principle
An answer can sound certain.
Your decision needs sources.
Can you trust an AI agent recommendation? OpenAI advises reviewing important facts because even polished responses can be incomplete or wrong. Read OpenAI's verification guidance. ChatGPT's web-search feature can supply citations; open those pages to see whether they support the specific recommendation. See OpenAI's web-search documentation.
Our editorial rule: save the answer's date, the exact claim and its original source. Check whether each source concerns the same professional, market and time period. Two websites repeating one brokerage biography still lead back to one first-party claim.
If an answer cites Minnesota Top Agents, attribute its 0–10 score to our editorial methodology. It is not a government rating, consumer review average or statewide MLS production ranking. Evidence confidence measures our documentation completeness, never agent quality.
Read the answer. Open the evidence.
Start with the answer in front of you. These are source-reading instructions. This guide does not rank services or agents by AI appearances. A name appearing in an answer does not establish the agent's quality or suitability.
Where did this recommendation come from?
Our reading notes
When ChatGPT uses web search, open the citations in its answer. Check whether the page names the same person and supports the specific claim.
Read ChatGPT source guidanceA fluent answer without a source leaves the claim unresolved. Record whether your answer used search.
Does the linked page support the claim?
Our reading notes
Look for the Sources button beneath or within the response when available. Open the relevant links and locate the actual supporting passage.
Read Gemini source guidanceSome responses have no sources. A related page is a lead to inspect; it does not automatically substantiate the recommendation.
Whose results does this citation describe?
Our reading notes
Open the numbered citations beside a claim and read the original page. Check whether its reviews or sales describe an individual, a team or an entire brokerage.
Read Perplexity source guidanceA citation number identifies a source. It is not an agent rank, and several links may repeat one first-party claim.
Is this in the overview or a nearby result?
Our reading notes
Open the supporting links in the overview, then check other Search results and the original sources. Record whether a name appeared in the overview itself.
Read Google AI Overviews source guidanceAn overview may not appear for your query. A nearby ordinary search result or advertisement is a separate placement.
Note the exact question, service and mode if shown, date, location settings, named person or team, and the URLs attached to the claim. A name mentioned in an answer and a website cited by that answer are different observations. Neither establishes agent quality or a platform endorsement.
Agent or team? See how to read real Minnesota examplesThe city evidence atlas
Explore who appears in our published city rankings—and how much of each roster has a traceable source record.
This measures our directory’s documentation, not Google or ChatGPT visibility, local sales, or agent quality.
12 published city rankings
64 distinct profiles across them
24 / 64 have exact sources
Dataset snapshot
Put the shortlist in place
Choose a city to update its agents and evidence below.
Minneapolis31 profiles · 10 with exact sources
Read Minneapolis's evidenceCity reference points, not agent locations or service boundaries. 2026 U.S. Census geography.
Every bar represents that city’s full roster.
Cities appear alphabetically. Select a city for its evidence breakdown; ↗ opens the full city guide.
Minneapolis selected. 10 of 31 profiles have exact source records. 21 do not have an exact source register recorded.
Urban Core / Minnesota
10 / 31
Categories overlap. An exact source supports specific recorded fields; it does not verify every claim on a profile.
10 of 31 profiles have exact source records (32.3%). The other 21 lack a stored field-level source register; that is a gap in our documentation, not an assessment of their work.
9 of 31 profiles have source coverage for market or specialty focus. That category can cover a broader service area or specialty; it does not establish recent transactions in Minneapolis. Ask for examples matching your location and property type.
23 of 31 profiles also appear in another published city roster. Repeated inclusion reflects overlapping editorial, home-base or service-area records, not independent endorsements or a measured share of local transactions.
Minneapolis / The names behind the city
The first 3 profiles in the city’s existing editorial order. The 0–10 score is our calculation; documentation confidence is separate. See the method ↗
9.1 / 10 editorial score
The Graham Smith Team · Keller Williams Integrity Lakes · Team
Graham Smith leads The Graham Smith Team at Keller Williams Integrity Lakes. Zillow displayed 119 team reviews at a 5.0 average, 80 team sales in the trailing 12 months and 1,419 total team sales when this profile was reviewed. Those figures describe the team—not Graham's individual production—and replace earlier unsourced cross-platform and career-volume totals.
A team model — you may work with team members, not always Graham directly
Full profile and limitationsProfile reviewed
Showing 2 source examples of 3 exact source URLs. Read the full source register for every recorded source.
Supports: team identity; Keller Williams affiliation claim; professional contact information; stated service model.
Supports: 119-review team aggregate; 5.0 team rating snapshot; 80 team sales in trailing 12 months; 1,419 total team-sales snapshot; $45K–$2.7M team price range; team attribution.
9.1 / 10 editorial score
Northeast Real Estate Group · Team
Jim Dropps is the broker-owner of the two-agent Northeast Real Estate Group roster. Zillow displayed 33 reviews at a 5.0 average, 16 sales in the trailing 12 months and 378 total recorded sales when checked September 20, 2026. His Zillow biography separately claims 350-plus homes sold and 12 built; those are first-party claims, not audited production totals, and this review did not capture a current Minnesota Commerce result.
Public review volume is modest next to high-volume Twin Cities teams, so the sample behind the rating is smaller
Full profile and limitationsProfile reviewed
Showing 2 source examples of 3 exact source URLs. Read the full source register for every recorded source.
Supports: broker-owner identity; Northeast Minneapolis brokerage base; two-agent brokerage roster; stated building and investment-property experience; profile-supplied Licensed Minnesota Broker credential and established-2012 claims; not current Commerce verification.
Supports: 33-review aggregate; 5.0 rating snapshot; 16 sales in trailing 12 months; 378 total-sales snapshot; $60K–$1M sold-price snapshot; platform-listed service areas.
8.9 / 10 editorial score
Edina Realty · Solo
A well-documented downtown Minneapolis specialist: Zillow displayed 159 reviews at 5.0 and 25 trailing-year sales, while Homes.com displayed 175 five-year closings with the largest neighborhood concentration in North Loop.
Heavily condo/downtown-weighted -- less of a fit for buyers set on a single-family home in the outer suburbs
Full profile and limitationsProfile reviewed
Showing 2 source examples of 3 exact source URLs. Read the full source register for every recorded source.
Supports: professional identity and Edina Realty brokerage affiliation; downtown Minneapolis office and residence claim; GRI and property-specialty designations; published contact information.
Supports: 159-review aggregate and 5.0 rating snapshot; 25 sales in the trailing 12 months and 1,784 total-sales platform snapshot; $70K to $2.8M price-range snapshot; 27-year experience claim; downtown Minneapolis service areas and specialties.
No payment changes inclusion or order. City membership can come from an editorial selection, a home-base record, or a recorded service area; it does not establish recent closings in that city. Official records and first-party claims retain their original scope and date.
One roster, overlapping places
191 city-list appearances come from 64 distinct profiles. The 127 appearances beyond each profile’s first listing reflect overlapping directory membership.
Five profiles with the widest footprints in our published city rosters, ordered by roster count; ties are alphabetical. These are not search or AI appearance counts, and they do not measure local production.
Chanhassen · Chaska · Deephaven · Eden Prairie · Excelsior · Minnetonka · Orono · Plymouth · Shorewood · Wayzata
Chanhassen · Chaska · Deephaven · Eden Prairie · Edina · Excelsior · Minnetonka · Orono · Shorewood · Wayzata
Deephaven · Eden Prairie · Edina · Excelsior · Minnetonka · Orono · Wayzata
Deephaven · Eden Prairie · Excelsior · Minnetonka · Orono · Plymouth · Wayzata
Chaska · Eden Prairie · Edina · Minneapolis · Minnetonka · Plymouth · Wayzata
02 / Read past the superlative
These are illustrative claim types, not statements about any named agent. Keep a claim unresolved until the source actually supports it.
Open or ask for
Find the named publisher, comparison group, criteria and review date behind the ranking.
What remains unproven
An unexplained first-place position does not establish who is best for your transaction. A publisher's editorial order belongs to that publisher.
Open or ask for
Open the Minnesota Commerce lookup. Match the legal name, license type, current status and brokerage information.
What remains unproven
A number copied into an AI answer or directory is only an identifier. If you cannot establish a match, leave the status unresolved.
Open or ask for
Open the actual review platform. Record the profile subject, rating, count and date you checked it.
What remains unproven
A team profile is not an individual review record. Keep platforms separate; do not add their totals or assume every review concerns the person you will hire.
Open or ask for
Find the original production source, period, geography and whether the figure covers a person or team, transactions or sides.
What remains unproven
Career and annual totals answer different questions. An agent or brokerage claim remains first-party evidence when an AI answer repeats it.
Open or ask for
Ask for recent, attributable examples involving your chosen area, property type and buyer or seller role.
What remains unproven
A service-area list is a starting point. It does not by itself document recent comparable work or guarantee availability.
For regulated status, start at the Minnesota Department of Commerce license lookup and choose Real Estate, Agent/Agency/Broker. For help with identity, enforcement and complaint distinctions, use our full Minnesota verification guide.
03 / A hypothetical example
Imagine an AI answer describes an unnamed agent as a top listing agent with hundreds of sales and excellent reviews. Its sales citation opens a team's career-total page; its review citation opens that same team's profile. Neither establishes the individual's recent listing work in your chosen city.
This example is invented to explain the method. It is not a finding about a real agent, team or ChatGPT response. Missing evidence is not evidence of poor performance.
“The linked sources describe a team. This person's recent local listing results, role in those transactions and current license status still need checking.”
“Which recent listings like mine did you personally handle, what was your role, and who would handle pricing, showings and negotiations for my sale?”
04 / Take this into your next chat
A better prompt than “Who is the best agent?” Replace the brackets with your own city, property type and buying or selling task. You can copy this prompt into your research chat. The output still needs the source checks above.
City, property type and transaction needs are enough to start. Keep private financial records, identification documents and confidential negotiation details out of this research prompt.
Help me research real-estate agents for [buying or selling] a [property type] in [Minnesota city]. Use current web sources where available. For each candidate, give the exact name, brokerage, source URL, source date if shown, and the specific claim that page supports. Distinguish individual from team evidence and first-party claims from independent platform records. Explain why the evidence is relevant to my transaction. Identify missing or conflicting facts. Link to the Minnesota Commerce license lookup and leave current status unresolved unless you have actually checked the matching official record. Do not invent a statewide ranking. Finish with three questions I should ask each candidate directly.
05 / From research to a real conversation
Search our referral verifier and inspect the profile's dated evidence. If there is no match, continue with official records; our roster is incomplete.
Check a recommended agentStart with the researched city guides or browse by transaction specialty. Read each profile's limitations before deciding whom to interview.
Browse agents by area and specialtyUse the same evidence categories for each person. Our comparison guide includes a private worksheet for recording what is supported and what remains unanswered.
Compare and choose a Minnesota agentBring the unanswered questions to a real conversation. Use the buyer's agent interview guide or the listing agent interview guide to discuss representation, responsibilities and agreement terms.
This guide does not establish a statewide ChatGPT winner. If your answer names someone first, inspect its original sources and selection criteria, then verify the person's identity and relevant experience. We do not present a generated answer as an OpenAI endorsement.
Use it to organize research around your chosen city, property type and buying or selling task. Ask for exact sources and unresolved questions. Confirm the information outside the chat and interview candidates about their actual role, availability and agreement terms.
Mark that claim unverified. Ask for the original page and read it yourself. A link to a homepage or a page about a different team member does not establish the specific claim. If evidence remains missing, ask the agent directly rather than filling the gap with an assumption.
Our roster is selective and incomplete. No match means we do not have a published profile; it says nothing about the person's license status or quality. Use the official Minnesota Commerce lookup and the same source checks for agents outside our directory.
The source register
First prepared October 3, 2026; published October 10, 2026; reviewed October 10, 2026. The sources below support the product guidance and official lookup path. The claim audit, hypothetical example and prompt are our editorial method; they are not a platform's agent-selection standard. This guide does not measure which agents ChatGPT recommends most often or establish an OpenAI endorsement. The same observation limit applies to Gemini, Perplexity and Google AI Overviews.
OpenAI · Checked October 10, 2026
OpenAI advises checking important names, dates and factual claims against current source material; a polished response can still be wrong or incomplete.
OpenAI · Checked October 10, 2026
Web search can provide current web information and citations in ChatGPT; this describes a product capability, not verification of an agent.
Minnesota Department of Commerce · Checked October 10, 2026
Commerce provides the official license lookup entry point, including Real Estate, Agent/Agency/Broker. This source record does not establish any named agent's status.
Google — Gemini Apps Help · Checked October 10, 2026
When supplied, Gemini sources and related links are accessible through a Sources button beneath or within a response. Not every response includes links.
Perplexity · Checked October 10, 2026
Perplexity describes numbered citations that link its answers to original sources.
Google — Search Help · Checked October 10, 2026
AI Overviews appear when Google's systems consider them helpful, can contain mistakes, and provide links for further checking. Google advises checking other Search results too.
Minnesota Top Agents is an independent digital publisher, not a brokerage, regulator or comprehensive background-check service. No payment may change inclusion, scores, numbered order, limitations or verdicts. Our directory does not contain a standardized private NorthstarMLS production feed. Read our sourcing and corrections policy.
Service marks identify the products discussed, with no affiliation or endorsement. OpenAI and ChatGPT are trademarks of OpenAI. Google and Gemini are trademarks of Google LLC. Perplexity marks belong to Perplexity AI. These editorial illustrations are not service screenshots. OpenAI brand information.
