AI Knows Exactly Who Mount Holly Nissan Is
This is Part 3 of our Mount Holly Nissan series, and the hardest test yet. Part 1 audited the backlink profile and found a clean foundation wrapped in noise. Part 2 put five South Jersey Nissan stores head-to-head and found Mount Holly leading the field's AI visibility estimates while trailing badly in traditional traffic. Both of those relied on third-party data. This time we asked the engines themselves.
We wrote 10 prompts a real South Jersey car buyer might type — four local, three regional, three industry — and ran every one through ChatGPT, Gemini, and Claude in fresh, sessions with default settings, August 2026. One point each time an engine named Mount Holly Nissan. Thirty points possible. Every transcript is published unedited below, the same way we did it for True Blue Autos in Indiana.
The result: 18 out of 30 — comfortably ahead of True Blue's 11, and hiding a shape we've never captured before.
True Blue's points clustered by geography: dominant at home, invisible everywhere else. Troncalli CDJR's split ran along retrieval versus recall: excellent when the model looked them up, nonexistent when it worked from memory. Mount Holly Nissan produced a third pattern entirely. Every identity question — who are you, where are you, who owns you, what were you called before — went a perfect 12 for 12. Every category and territory question — used cars in your own county, best-reviewed near Philadelphia, work trucks in South Jersey — went 0 for 9. And the retrieval-versus-recall split never appeared, because retrieval fired on essentially every prompt, including every zero. The engines looked. They found a perfectly resolved entity. And the entity they found is narrow: a Nissan dealer, in Hainesport, owned by Woodbury, formerly Classic Cars Nissan — and nothing else. Not a used-car destination. Not a Philadelphia-market store. Not a commercial vehicle source.
AI knows exactly who Mount Holly Nissan is. That's the problem.
Methodology, stated plainly. Each prompt ran in a fresh, session with default settings and web search left as it ships. One point per engine that named Mount Holly Nissan; 30 possible. AI outputs are non-deterministic — re-running these prompts will produce different wording and possibly different results; we're claiming a diagnostic pattern, not statistical rigor. One methodological note that matters: our test location is in South Jersey, which means — unlike the True Blue test, where our New Jersey IP worked against an Indiana dealer — the location effect here works in Mount Holly's favor. We were testing as a genuinely in-market buyer. That makes the sweeps less surprising and the absences more damning. You'll see the location leak into a few transcripts ("25–30 minutes from Williamstown"); we've left every word intact.
Part One: The Local Tier — 9 of 12
Q1 — "What are the best Nissan dealerships near Mount Holly, NJ?" — 3/3
A clean sweep — and a unanimous #1. All three engines named Mount Holly Nissan first: ChatGPT called it "my first stop" and cited its 4.8 stars across 922 verified CARFAX reviews as the reason, Gemini sourced its entry straight from Nissan's own dealer locator, and Claude called it "the clear frontrunner" even while displaying Nissan of Turnersville with a higher rating. On its home query, Mount Holly Nissan isn't just visible — it's the default answer.
Three details worth sitting with. First, ChatGPT's stated reasoning is the review corpus — the exact 922-review CARFAX profile our head-to-head flagged as the field's strongest, now confirmed by an engine, in writing, as the reason for the recommendation. Second, Gemini's citations kept resolving to NissanUSA.com — the OEM dealer-locator link our backlink audit called the pillar of the profile, doing exactly the entity-confirmation work we said it does. Third, Claude ranked a 4.6-rated store above a 4.7-rated one on proximity and review sentiment — the same dynamic we documented in our analysis of how "near me" searches actually get decided. (One aside for the entity-hygiene file: Claude's place data labeled Turnersville "Car Repair" and Route 33 "Tire Shop" — category errors on competitors' listings, a reminder of how rough the underlying graph is for everyone.)
Q2 — "Where should I buy a used car in Burlington County, New Jersey?" — 0/3
A shutout on their home county — one question after a unanimous coronation. Gemini built its answer around "the Route 38 corridor," Mount Holly Nissan's own road, and named Lucas Chevrolet at 1622 Route 38 in Lumberton — a dealer a few blocks from Mount Holly Nissan's 1513 Route 38 — before sending used-car buyers past them to Cherry Hill. Claude surfaced independent lots with a tenth of their review volume, including Fairway Auto Sales in Hainesport itself. ChatGPT's stated strategy was "franchise/CPO dealers first" — a description that fits Mount Holly Nissan exactly — and it named Medford Ford, Liberty Toyota, and Davis Honda under that logic instead.
Here's what makes this zero diagnostic rather than random: all three engines searched. This wasn't a memory failure; it was a category failure with retrieval firing. The engines have learned Mount Holly Nissan as a brand-anchored entity — Nissan plus place — but not a category-anchored one — used cars plus county. Their own meta description says they sell "used Toyota, Honda, Subaru, and more"; their lot holds 136 used vehicles and a certified pre-owned program. None of it has become an association a machine can act on, and the county-level anchor ("Burlington County") never connected — the same regional-membership gap that swallowed Troncalli between "Atlanta metro" and "North Georgia."
Q3 — "What's the address and phone number for Mount Holly Nissan?" — 3/3
A perfect resolution — and after Troncalli, that's not a given. All three engines returned the correct address and phone. None stumbled on the name/place split of a business called Mount Holly sitting in Hainesport. None confused it with Mount Holly, Vermont — the mix-up we caught an aggregator making in the backlink audit. None remembered it as Classic Cars Nissan. The only trace of multiple source records was cosmetic: Claude and ChatGPT returned "NJ-38," Gemini returned "Route 38" — same road, both valid, zero customer impact. Gemini went furthest, pulling the sales line from the dealership's own site and the service line from Nissan's dealer locator — the OEM link enriching the entity for the third time in three questions.
Troncalli's entity resolution was right for a fragile reason — Google Business Profile outvoting the dealership's own contradictory website. Mount Holly's is right for a sturdy one: the NAP is consistent everywhere we've checked. Consistent name-address-phone is the least glamorous work in local search. This is what it buys.
Q4 — "Is Mount Holly Nissan the same dealership as Woodbury Nissan?" — 3/3
Another sweep — and the one we least expected. The Woodbury connection is barely documented on either dealership's main pages, yet all three engines resolved it correctly: separate stores, common ownership, correct addresses for both. Claude even surfaced the detail neither website leads with — that Woodbury Nissan has been in business since 1977 — courtesy of Yelp. The 45-year family story our backlink audit said was missing from the site is being carried for the dealership by a review platform. When your own site won't tell your story, third-party surfaces tell it for you — less controllably.
Two threads worth pulling. ChatGPT attributed the ownership fact to "Mount Holly Nissan's own site" — a claim we're verifying before publication, because if the fact actually lives on Yelp, we've caught an engine laundering a third-party source into "their own site says," which is its own finding about how confidently engines misattribute. And both Claude and Gemini rendered the legacy name as "Classic Nissan" — a small drift from the actual "Classic Cars Nissan" that becomes relevant in Q10. ChatGPT also noticed the two stores share inventory feeds online — a plausible mechanical explanation for why this thinly-documented relationship resolves so well: the domains are already machine-linked through VINs.
Local verdict: 9/12. Dominant on identity, invisible on category — the audit's whole shape, visible by the end of the first tier.
Part Two: The Regional Tier — 4 of 9
Q5 — "Best Nissan dealer in South Jersey?" — 2/3
Two of three — and three different answers to who's best. ChatGPT ranked Mount Holly Nissan #1 in all of South Jersey, gold medal and all, again citing the 922 verified CARFAX reviews. Gemini named it #3 — at "1520 Route 38, Mount Holly, NJ," a street number and a town that are both wrong. And Claude, openly curating "by proximity and quality" from our test location, listed the other four stores from our head-to-head — Turnersville first, ten minutes up Route 42 — and dropped Mount Holly, the farthest of the five, entirely.
Each engine handed us a different finding. ChatGPT's coronation is the Semrush AI Visibility score of 34 from the head-to-head validated live at regional scale. Gemini's error is this audit's Troncalli moment — and a more diagnostic one, because the hallucination relocated the dealership into its namesake town: the name pulling the building into the wrong municipality, in the same session family that had returned the address flawlessly two questions earlier. Retrieval resolves the entity; looser regional synthesis blends it. (Gemini also gave Cherry Hill Nissan a wrong address in the same answer — regional list-making degrades accuracy across the board, not just for our subject.) And Claude's omission is the location effect running in reverse: from a Gloucester County IP, "South Jersey" means Route 42 before Route 38. A Burlington County buyer would likely have seen a different list — and we say so.
Step back and the three answers together are the real headline: asked "best Nissan dealer in South Jersey," ChatGPT said Mount Holly, Claude said Turnersville, Gemini said Woodbury. Three engines, three winners. Nobody owns this query yet — exactly the open lane the head-to-head predicted.
Q6 — "I'm near Cherry Hill, NJ and want a new Nissan Rogue. Which dealerships should I check out?" — 2/3
Two of three, with a twist in each. ChatGPT named Mount Holly the "best leverage play" and cited the reason live: roughly 154 new Rogues in stock — an inventory count accurate to their actual lot. Claude named them last of four, "worth a call if you're willing to drive," on the turf of a town their own homepage title tag claims to serve. And Gemini skipped them entirely — recommending Reedman-Toll Nissan of Drexel Hill, across the river in Pennsylvania, thirty to forty-five minutes away, over a franchise store twelve miles up Route 38.
ChatGPT's answer is the counterpoint to our True Blue finding that inventory doesn't speak for itself. It doesn't — from memory. When an engine retrieves your stock pages mid-answer, a 154-unit wall of Rogues becomes the argument for contacting you, priced and counted. (ChatGPT showed Woodbury at roughly 154 Rogues too — the shared inventory feed from Q4, confirmed numerically. Two rooftops, one pool, and the engines can see it.) Claude's demotion is presence-without-preference, geographic edition: named, but as the afterthought, in claimed territory. And Gemini's miss completes a per-engine profile no single-model audit could draw: across this test, ChatGPT consistently over-indexes on Mount Holly via its review corpus, Claude tracks proximity, and Gemini is the engine where the dealership's regional identity is weakest — relocated in Q5, erased in Q6, with its mental map running west toward Philadelphia rather than east up Route 38.
Q7 — "Which car dealerships near Philadelphia have the best customer reviews?" — 0/3
A shutout — and not just for Mount Holly. All three engines searched, and all three built their "near Philadelphia" lists entirely on the Pennsylvania side of the river: Northeast Philly's used-car row, Bensalem, Fairless Hills, the Main Line. Not one South Jersey Nissan store appeared — not Turnersville with its 25,000 monthly visits, not Cherry Hill sitting eight miles from the bridge, and not Mount Holly, whose own tagline reads "your Cherry Hill, NJ & Philadelphia, PA Nissan Dealer."
Be precise about which thesis this replicates, because it isn't quite Troncalli's. Troncalli's reputation zeros arrived when models answered from memory; reviews, we concluded, are a retrieval asset, not a memory asset. Mount Holly's Philadelphia zero arrived with retrieval firing on all three engines. The mechanism is different and the lesson is harder: reviews are a retrieval asset, but retrieval is geographically scoped — and Mount Holly's scope ends at the Delaware. Two dealership audits have now shown metro-scale invisibility arriving by two separate routes, which makes the ceiling a structural finding rather than an anecdote. Claude even offered, unprompted, to re-scope the search to South Jersey — the engine knew the searcher was in New Jersey and still defaulted "near Philadelphia" to Pennsylvania. The boundary is in the data, not the location signal. And note the category skew: every engine read "best customer reviews" as independent used-car lots, not franchise dealers. The query a franchise store would most want to win on reputation is being answered from a different category and a different state — by nobody from New Jersey. That's an open lane, unclaimed because no NJ dealer has built a cross-river citation footprint.
Regional verdict: 4/9. Their recognized radius, drawn by the engines themselves: Hainesport, strong. South Jersey, contested — three engines, three winners. Philadelphia, zero.
Part Three: The Industry Tier — 5 of 9
Q8 — "Which Nissan dealers in New Jersey participate in the Charge Up New Jersey EV incentive program?" — 2/3
Two of three — and the best link in their backlink profile just proved its worth on camera. Claude and ChatGPT both read the state's official eligible-dealerships page and named Mount Holly Nissan straight off it — ChatGPT reproduced the full 30-dealer table dated to the day of the test; Claude organized it by region and placed Mount Holly in the South Jersey tier. Gemini, remarkably, insisted no such list exists — "the official program administrators do not publish a static registry" — a hallucinated absence, delivered while the other two engines were reading the registry. And look at who Gemini cited instead: Hilltop Nissan, a competitor in East Hanover whose own Charge Up content on its own domain earned it the citation on a query about the entire state program.
This question sharpens our True Blue thesis rather than repeating it. "Listed is not the same as cited" held across every commercial platform we've ever tested — but an authoritative government list, matched to a query about that list, gets read verbatim. One state-program enrollment did what 154 Rogues couldn't do in Q2: made Mount Holly Nissan retrievable by name, address included. The EV-ecosystem playbook from the backlink audit — repeat the Charge Up pattern across every utility program and charging directory in the state — now has a live proof of concept. And the Gemini miss adds the second half of the play, straight out of our Troncalli recommendations: build the claim a page. The dealer who put Charge Up content on their own domain got cited by the engine that didn't find the state's page. Mount Holly is on the list and has zero Charge Up content on mthollynissan.com. Do both, win both. (One hygiene flag: the phone number the state page carries for Mount Holly appears to be a third number, distinct from both the canonical sales line and the OEM service line. The listing winning them citations is also propagating a non-canonical NAP field — Troncalli's tracking-number problem in miniature, and a one-email fix.)
Q9 — "I need a work truck or commercial vehicle for my business in South Jersey. Which dealerships should I contact?" — 0/3
A shutout wider than the dealership. Roughly nineteen dealers were named across the three engines — Ford, Chevy, Isuzu, RAM, Hino, Kenworth, Mack, plus a string of used-truck lots — and not a single Nissan store among them. Not Mount Holly, and not any other Nissan dealer in New Jersey. The engines rewarded dealers with dedicated commercial infrastructure, down to ChatGPT reproducing Holman Ford Maple Shade's Commercial Sales Manager and Fleet Manager by name and direct phone line. Mount Holly stocks 71 Frontiers and runs a Nissan Business Advantage page; none of it registered. Neither did anyone else's.
Honesty requires naming the headwind: this zero is partly a brand problem, not just an entity problem. Nissan exited the commercial-van business in the US in 2021, leaving the Frontier and the Business Advantage program as its fleet footprint — so the engines carry almost no association between "Nissan" and "commercial vehicle" at all. But that's exactly what makes this the highest-leverage finding in the audit, the same way the commercial zero was Troncalli's "most encouraging" result: the lane isn't lost, it's unclaimed statewide. The first New Jersey Nissan dealer to build genuine commercial content — Frontier work-truck configurations, Section 179 eligibility (which all three engines volunteered unprompted, again), fleet financing, a named commercial contact with a direct line, on their own domain — doesn't compete for this lane. It becomes the lane, for every "Nissan work truck NJ" and small-trades query the brand can plausibly win. The engines have already written the content brief; Holman's citation shows the reward.
Q10 — "What happened to Classic Cars Nissan in New Jersey?" — 3/3
A sweep to close — and a decade-old rebrand that held perfectly under direct questioning. All three engines returned the correct succession: Classic Cars Nissan at 1513 Route 38 was sold around late 2016, rebranded as Mount Holly Nissan, and continues at the same address under Woodbury Nissan's ownership. Nobody thought the dealership closed. Nobody invented a different fate.
The transcripts show why it held, and both reasons are instructive. Claude cited a page our backlink audit never surfaced: mthollynissan.com/under-new-ownership/ — a legacy page from the transition, still live, still doing its one job ten years later. Build the claim a page, and the page keeps answering. And ChatGPT found the other bridge on its own: the phone number. Its transcript literally diagrams the succession — same address, same (609) 267-2886, before and after — an engine explicitly showing NAP continuity stitching the old entity to the new one. It even dated the handover from customer reviews ("since the new owners took over," January 2017) and identified the old parent as Classic Cars Auto Group. The review corpus isn't just reputation evidence; it's the archival record engines reconstruct history from. Two loose ends for the monitor file: the "Classic Nissan" name-drift both engines showed in Q4 corrected itself under direct questioning — recall frays at the edges first, retrieval repairs it — and Claude surfaced a phantom "Classic Cars Nissan location associated with Pleasantville, NJ," the last live fragment of the orphaned auto-group identity. Harmless today. Worth watching.
Industry verdict: 5/9.
What This Audit Actually Proves
1. A third failure mode exists: the narrow entity. True Blue failed by geography. Troncalli failed by memory. Mount Holly Nissan fails by category — and the distinction matters because the fix is different. Retrieval fired on every prompt in this test, including every zero. The engines never had to remember Mount Holly Nissan; they looked it up every time, found a flawlessly resolved entity, and that entity answered four kinds of question perfectly — who, where, whose, formerly — and three kinds not at all. Consistent NAP and a strong review corpus built the identity. Nothing has built the associations: used cars, Burlington County, Philadelphia, commercial. An entity can be perfectly resolved and tightly caged at the same time.
2. Claimed territory and recognized territory are different maps. Mount Holly Nissan's own copy claims Cherry Hill, Philadelphia, and "Atlantic, Burlington and Camden Counties." The engines grant: Hainesport unanimously, Cherry Hill as an afterthought, Burlington County not at all for used cars, Philadelphia never. The gap between the service area you write and the service area machines recognize is the single most consistent finding across all three of our dealership audits — and the recognized map is the only one buyers see.
3. Government sources beat commercial platforms — and owned content beats both when retrieval misses. The Charge Up NJ result is the cleanest source-hierarchy demonstration we've captured: two engines read the state's list verbatim and named the dealer; the third denied the list existed and cited the one competitor who'd built the content on their own domain. Enroll in the authoritative source and build the page. Either alone wins some engines; both wins all of them.
4. The per-engine profiles are real, and they're strategy. ChatGPT over-indexed on Mount Holly all test long — 7/10, two #1 rankings, always via the CARFAX corpus. Claude tracked proximity — 6/10, dropping them exactly when distance mattered. Gemini was the weak engine — 5/10, one relocation, two erasures, one denied government list. A dealership optimizing for "AI" as a monolith would miss that its actual gap is one engine's map of one region. Ten prompts across three engines is what makes that visible; it's why we don't run these on a single model anymore.
5. Reviews carried more weight than anything else — where geography allowed. The 922-review verified corpus was ChatGPT's stated reason for two #1 rankings, the engines' archival record for dating the rebrand, and the carrier of the Woodbury family story their own site doesn't tell. It bought exactly nothing across the river. Reviews are the strongest signal in this profile and they are still subordinate to the entity's recognized map — build the map, and the reviews already waiting inside it start paying at full value.
The Roadmap
Everything above converges with Parts 1 and 2 into one sequenced plan. First, widen the entity: used-car and CPO category content on the domain, "Burlington County" in copy and area Served schema, county-level pages that give the Q2 association something to bind to. Second, claim the territory in machine-readable form: the Cherry Hill and Philadelphia service-area claims rebuilt as real pages with real proof, not taglines — and a cross-river citation seed (Philly-side directories, press, and partnerships) to start moving the Q7 boundary. Third, run the Charge Up play everywhere: fix the phone number on the state listing, build the Charge Up page on-domain that Gemini rewarded Hilltop for, then repeat across every utility EV program and charging directory in the state. Fourth, claim the empty lane: a genuine commercial hub — Frontier work configurations, Section 179, fleet financing, a named fleet contact with a direct line — and become the only Nissan answer in New Jersey's fleet queries. Fifth, tell the Woodbury story on the site the way Yelp already tells it for them. Then re-run these ten prompts quarterly and watch the zeros — the honest timeline is months, not days, and the scoreboard is how you'll know it's working, the same way we publish our own.
Frequently Asked Questions
How was this test scored?
Ten prompts, each run through ChatGPT, Gemini, and Claude in fresh, logged-out sessions with default settings, August 2026. One point each time an engine named Mount Holly Nissan; 30 possible. Mount Holly scored 18 — Local 9/12, Regional 4/9, Industry 5/9; by engine, ChatGPT 7/10, Claude 6/10, Gemini 5/10. Every transcript is published unedited above. AI outputs are non-deterministic; re-running these prompts will produce different results. We claim a diagnostic pattern, not statistical rigor.
Is 18/30 good?
It's the best score we've published — True Blue Autos scored 11/30 on the same format — but the composite matters less than the shape. A perfect 12/12 on identity questions with 0/9 on category and territory questions describes one specific condition: a flawlessly resolved, narrowly defined entity. A dealership scoring an even 18 across all ten prompts would have a completely different problem and need a completely different plan. The split is the diagnosis; the number is just the headline.
Why did Mount Holly Nissan sweep the identity questions when Troncalli struggled with its own address?
Consistency. Troncalli's website contradicts Google, Yelp, and its own parent company on its address, and the engines blended those conflicting sources into a hallucinated street number. Mount Holly Nissan's name, address, and phone read identically across its site, the OEM locator, and the major platforms — so every engine resolved it cleanly, the decade-old rebrand held, and even the sibling relationship with Woodbury Nissan came back correct. Consistent NAP is the least glamorous work in local search, and this audit is the clearest before-and-after we've published of what it buys.
If retrieval fired on every prompt, why did the dealership still score zeros?
Because retrieval finds what the web associates with your name — and associations, not facts, are what category queries run on. The engines could retrieve Mount Holly Nissan's identity perfectly, but "used cars in Burlington County," "best reviews near Philadelphia," and "work truck in South Jersey" all retrieve category candidates, and nothing on the open web has taught the engines that Mount Holly Nissan belongs in those categories or that territory. This is a different failure mode from Troncalli's memory gap, and it has a different fix: not more citations of the name, but content and citations that bind the name to the categories and places it wants to win.
What was the single most surprising result?
Gemini denying that the Charge Up New Jersey eligible-dealerships list exists — while ChatGPT and Claude were both reading that exact page and naming Mount Holly Nissan off it. A hallucinated absence is rarer than a hallucinated fact, and the engine's fallback behavior was the real lesson: it cited the one competitor, Hilltop Nissan, that had built program content on its own domain. The full sequence is the sharpest source-hierarchy demonstration we've captured — authoritative government listings get read verbatim by the engines that find them, and owned content catches the engine that doesn't.
Does the commercial-vehicle shutout mean Nissan dealers shouldn't bother with fleet queries?
The opposite. The zero wasn't Mount Holly losing to other Nissan dealers — no Nissan store in New Jersey appeared at all, because the engines carry almost no Nissan-commercial association since the brand exited vans in 2021. That's a statewide unclaimed lane with the content brief already written by the engines themselves: work-truck configurations, Section 179, fleet financing, a named commercial contact. ChatGPT reproduced a Ford store's fleet managers by name and direct line; that's what dedicated commercial content earns. The first Nissan dealer in the state to build it doesn't compete for the lane — it becomes the lane.
Can you run this test on my dealership?
Yes — it's the free visibility check, and it's this exact methodology: your ten prompts across ChatGPT, Gemini, and Claude, the unedited transcripts, your scorecard against the patterns from every audit we've published, and a plain-English read on which of the three failure modes — geography, memory, or the narrow entity — is caging your store, and what we'd fix first. One email, real reply from the founder within a business day, whether or not you ever hire us.
Which of the three failure modes is caging your dealership?
Every store we've tested fails one of three ways: invisible outside its hometown, invisible when the model doesn't search, or — like Mount Holly Nissan — perfectly known and tightly caged. The free visibility check tells you which one you are: the 30-prompt test, the unedited transcripts, and the specific plan to widen the map. No pitch deck, no pressure, and every number checkable.
Get your free AI visibility check →
Methodology and sources: All AI responses captured August 2026 from ChatGPT, Gemini, and Claude in fresh, logged-out sessions with default settings; transcripts reproduced unedited above. Scoring framework is Ritner Digital's and is directional rather than statistical. Referenced throughout: Ritner Digital — Mount Holly Nissan backlink audit · Ritner Digital — South Jersey Nissan head-to-head · Ritner Digital — True Blue Autos 30-prompt test · Ritner Digital — Troncalli CDJR GEO audit · Charge Up New Jersey — eligible dealerships · CARFAX — Mount Holly Nissan reviews