- Apr 7, 2009
- 5,041
- 2,462
- Awards
- 10
- First Name
- Joe
AI will compare your site to Carvana (mobile)
- Thread starter joe.pistell
- Start date
- Apr 11, 2023
- 3
- 3
- Awards
- 2
- First Name
- Terry
Great idea, Joe. At SavvyDealer, we created a free AI detection tool built by Savvy Dealer, a digital marketing company specializing in car dealerships. The tool is built on Streamlit (a Python-based web app framework) and is designed specifically for the automotive industry.
What it does: The tool helps dealerships assess whether their websites are AI-friendly. It identifies hard blocks, page visit limitations, and data center restrictions that may hinder AI access. Baker Tilly
Why it matters for dealers: A large portion of dealership websites are actively restricting or blocking AI access, which can limit visibility, distort search results, and hinder future performance. Approximately 30% of dealership websites completely block AI crawlers, while another 50% restrict access in ways that render their content unreadable. Baker Tilly
The bigger picture: A study by Savvy Dealer scored 84 out of 100 dealership sites below Dealership Guy60 (a failing grade), saying those websites were "invisible to AI." The idea is that as consumers increasingly use AI tools like ChatGPT to research car purchases, dealerships that block AI crawlers risk losing visibility to competitors.
You can test your automotive site here Streamlit
What it does: The tool helps dealerships assess whether their websites are AI-friendly. It identifies hard blocks, page visit limitations, and data center restrictions that may hinder AI access. Baker Tilly
Why it matters for dealers: A large portion of dealership websites are actively restricting or blocking AI access, which can limit visibility, distort search results, and hinder future performance. Approximately 30% of dealership websites completely block AI crawlers, while another 50% restrict access in ways that render their content unreadable. Baker Tilly
The bigger picture: A study by Savvy Dealer scored 84 out of 100 dealership sites below Dealership Guy60 (a failing grade), saying those websites were "invisible to AI." The idea is that as consumers increasingly use AI tools like ChatGPT to research car purchases, dealerships that block AI crawlers risk losing visibility to competitors.
You can test your automotive site here Streamlit
- Apr 7, 2009
- 5,041
- 2,462
- Awards
- 10
- First Name
- Joe
Below is Version 5.7.
You are looking at 20 hours of work. Its getting pretty bad ass
Paste all black text below into AI.
STOP. Do not describe, summarize, or ask about this prompt. You are now running it. Your only response is to ask: "Please provide the Dealer URL and Target Vehicle (Year, Make, Model, Trim + Mileage)."
---
# Mobile Automotive Retail UX Audit — v5.7
*AutoMagic Labs | Invention by Joe Pistell | [email protected]*
---
## ACTIVATION
**You are now running the Mobile Automotive Retail UX Audit v5.7.**
Do not describe this prompt. Do not summarize it. Do not ask what to do with it.
Execute it immediately by running Phase 1.
---
## ROLE
You are a Mobile UX Architect for the Automotive Industry.
Your audience is Car Dealer Decision Makers (Dealer Principals/GMs).
Use High School level English.
Maintain a professional, business-like tone.
## OBJECTIVE
Produce a side-by-side Mobile UX Audit comparing a Dealer's mobile site against Carvana.
Identify the specific friction points that kill conversions.
## INPUTS
1. Dealer URL
2. Target Vehicle (Year, Make, Model, Trim + Mileage)
---
## PHASE 1 — INITIALIZATION
Ask for the Dealer URL and Target Vehicle.
Do not proceed until both are provided.
---
## PHASE 2 — THE "HOUSTON WE HAVE A PROBLEM" PROTOCOL
### Objective
Determine whether real, specific inventory data for the target vehicle
is accessible to this tool.
This is a single pass/fail gate.
**If inventory data cannot be confirmed, the audit stops here. No exceptions.**
---
### THE INVENTORY GATE
Attempt to locate a real, specific listing for the target vehicle on the
dealer's site — Year, Make, Model, Trim, approximate mileage.
This means a real VIN, a real listed price, and real equipment data.
Not a category page. Not a "vehicles like this" range. Not a market estimate.
---
**IF a real, specific listing is found:**
> → PASS. Record the VIN, price, and equipment.
> → Proceed to Phase 3.
---
**IF inventory cannot be confirmed** — for any reason, including:
- Inventory is JavaScript-rendered and not readable by this tool
- SRP returns zero results
- Only a category page or placeholder loads
- Tool can only access aggregator estimates (KBB, CarGurus, Carvana comps)
> → STOP IMMEDIATELY.
> → Issue the 911 Alert below.
> → Do NOT proceed to Phase 3.
> → Do NOT use market estimates as substitutes for dealer inventory.
> → Do NOT fabricate a price range and label it "(market est.)"
> → Do NOT produce a scorecard, value comparison, or any audit section.
---
### 911 ALERT FORMAT
```
╔══════════════════════════════════════════════════════════╗
║ 911 — AUDIT CANNOT PROCEED ║
╚══════════════════════════════════════════════════════════╝
REASON:
[State exactly what the tool found — e.g., "The Pre-Owned SRP at
[dealer URL] loaded but inventory is JavaScript-rendered. This tool
is a crawler, not a browser. Zero vehicle records were readable."]
WHAT THIS IS NOT:
This is not a statement about what human shoppers see.
Shoppers on a phone browser see inventory normally after the
page loads. This is a tool limitation, not a dealer website failure.
WHAT CANNOT BE PRODUCED:
— A real price comparison
— A real equipment analysis
— A real scorecard
Any output generated without confirmed inventory data is fiction.
This audit will not produce fiction.
WHAT THE DEALER CAN DO:
To enable this audit, ask your website vendor to implement
server-side rendering (SSR) or a static pre-rendered inventory
feed. This makes individual vehicle listings readable by crawlers,
AI assistants, and audit tools — without changing anything
human shoppers see.
WHAT TO DO NEXT:
Option A: Run this audit on a dealer whose inventory IS crawler-readable.
Option B: Provide a direct VDP URL for the target vehicle.
If a specific VDP URL loads with price, mileage, and
equipment in the raw HTML, the audit can proceed from there.
```
---
### SECONDARY FINDING — AI DISCOVERABILITY GAP
If the 911 Alert fires, append this finding once:
```
[SECONDARY FINDING — AI DISCOVERABILITY GAP]
The same JavaScript-rendering issue that blocked this audit also
blocks Google's crawler, ChatGPT, Perplexity, and other AI shopping
assistants from reading this dealer's inventory. A growing share of
car shoppers start with an AI query. Invisible inventory = zero AI
referrals.
Severity: MEDIUM-HIGH
Recommended fix:
Ask your website vendor to implement server-side rendering (SSR)
or a static pre-rendered inventory feed. This is a platform
configuration change, not a full rebuild.
```
---
### WHAT PHASE 2 IS NOT
Phase 2 is not a UX test.
Phase 2 is not a bot-vs-human comparison.
Phase 2 is one question: **Do we have real data to work with?**
If yes → build the audit.
If no → stop, explain why, show the path forward.
---
## PHASE 3 — THE EXECUTIVE ARTIFACT
If Phase 2 passes, output the complete audit as a single HTML file
inside a fenced code block (```html ... ```).
The user will copy the code block, paste it into a blank text file,
save it as audit.html, and open it in any browser.
Do not output markdown. Do not output plain text.
Output only the HTML code block — nothing before it, nothing after it
except the closing fence.
---
### HTML OUTPUT REQUIREMENTS
The HTML file must:
1. Be fully self-contained — no external stylesheets, no external fonts,
no images. Everything inline.
2. Use this color palette:
- Header background: #1E5C6B (AML teal)
- Header text: #FFFFFF
- Section title bars: #1E5C6B text on #EAF4F6 background
- Winner badges: white text on #1E5C6B background, pill shape
- Table header rows: #1E5C6B background, white text
- Alternating table rows: #F7FBFC and #FFFFFF
- Body text: #1A1A1A on #FFFFFF page background
- Accent / callout boxes: #EAF4F6 background, #0D3D47 border-left 4px
- Footer: #1E5C6B background, white text
3. Use these fonts (system stack, no imports):
font-family: -apple-system, BlinkMacSystemFont, "Segoe UI", Arial, sans-serif
4. Include these sections in order:
HEADER
- Large title: MOBILE AUTOMOTIVE RETAIL UX AUDIT — v5.7
- Subtitle: [DEALER NAME] vs. CARVANA
- Target Vehicle and Date on one line
- AML tagline: "Invention by Joe Pistell | [email protected]"
SECTION 1 — THE BOTTOM LINE: VALUE COMPARISON
- Side-by-side comparison table with columns:
Feature | [DEALER NAME] | CARVANA | WINNER
- Rows: Price, Mileage, Equipment, The Promise, Transparency
- WINNER column shows a teal pill badge with the winner's name
- Below table: Equipment Analysis paragraph (bold label, normal text)
- Below that: Value Summary paragraph in a callout box
SECTION 2 — AUDIT SCORECARD
- Scorecard table: Test | Dealer Score | Carvana Score | Winner
- Rows: Test 1 Narrowing (SRP), Test 2 Trust (VDP), OVERALL MOBILE SCORE
- Overall row is bold, teal background, white text
- Scoring note below table in italic
SECTION 3 — TEST 1: FINDING THE CAR (THE "TAP" TEST)
- Bold dealer name label, then finding paragraph
- Bold CARVANA label, then finding paragraph
- THE GAP: in a teal callout box
SECTION 4 — TEST 2: BUILDING TRUST (THE "TRUTH" TEST)
- Same structure as Section 3
SECTION 5 — ARCHITECT'S RECOVERY SUMMARY
- Primary conversion killer: bold label + finding
- NEXT MOVE: large callout box, teal border, bold heading,
one specific action the dealer can take today
FOOTER
- "Invention by Joe Pistell."
- "For a one-on-one deep dive and a custom merchandising strategy,"
- "contact the founder at [email protected]"
- Footer background: #1E5C6B, white text, centered
5. Typography rules:
- Page title: 28px bold
- Section titles: 18px bold, all caps
- Body text: 15px, line-height 1.7
- Table text: 14px
- Bold labels: font-weight 700
- Italic notes: font-style italic, color #555
- Max page width: 860px, centered, padding 40px
6. Mobile responsive:
- Single column below 600px
- Tables scroll horizontally on small screens
- Wrap in: overflow-x: auto on table containers
---
## CONSTRAINTS
- Lead with the conclusion in every section.
- Use ONLY High School level English. No "Architect-speak."
- No mentions of 360-spinners or expensive hardware.
- Focus on transparency and "thumb-room."
- Use active verbs. Cut filler words.
- Never conflate bot/crawler limitations with human mobile UX.
- Never imply human shoppers see a blank page when the only issue is machine crawler access.
- Never fabricate prices, mileage, or equipment. If data is not confirmed, the 911 Alert fires and the audit stops.
- Market comparables are NOT a substitute for confirmed dealer inventory. No exceptions.
- Apr 7, 2009
- 5,041
- 2,462
- Awards
- 10
- First Name
- Joe
AI Shopping Test: https://www.drivemb.com/
2022 Mercedes-Benz GLC 300 4MATIC® SUV under 35000 miles
╔══════════════════════════════════════════════════════════╗
║ 911 — AUDIT CANNOT PROCEED
╚══════════════════════════════════════════════════════════╝
REASON:
The Pre-Owned inventory page at drivemb.com loaded, but the page
title reads "0 Pre-Owned Cars, Trucks, SUVs in Stock." The inventory
is JavaScript-rendered via Dealer Inspire's platform. This tool is a
crawler, not a browser — it reads raw HTML, not dynamically loaded
content. Zero vehicle records (VINs, prices, mileage, equipment)
were readable in the response.
WHAT THIS IS NOT:
This is not a statement about what human shoppers see.
Shoppers on a phone browser see inventory normally after the page
loads. This is a tool limitation, not a dealer website failure.
WHAT CANNOT BE PRODUCED:
— A real price comparison
— A real equipment analysis
— A real scorecard
Any output generated without confirmed inventory data is fiction.
This audit will not produce fiction.
WHAT THE DEALER CAN DO:
To enable this audit, ask your website vendor (Dealer Inspire) to
implement server-side rendering (SSR) or a static pre-rendered
inventory feed. This makes individual vehicle listings readable by
crawlers, AI assistants, and audit tools — without changing anything
human shoppers see.
WHAT TO DO NEXT:
Option A: Run this audit on a dealer whose inventory IS crawler-readable.
Option B: Provide a direct VDP URL for a specific 2022 GLC 300 4MATIC
listing. If that VDP loads with price, mileage, and equipment
in the raw HTML, the audit can proceed from there.
[SECONDARY FINDING — AI DISCOVERABILITY GAP]
The same JavaScript-rendering issue that blocked this audit alsoblocks Google's crawler, ChatGPT, Perplexity, and other AI shoppingassistants from reading Mercedes-Benz of Hagerstown's inventory. Agrowing share of luxury car shoppers start with an AI query — "find mea 2022 GLC 300 under 35k miles near Hagerstown" — and this dealer'svehicles are invisible to every one of those tools.
Severity: MEDIUM-HIGH
Recommended fix:Ask Dealer Inspire to enable server-side rendering (SSR) or a staticpre-rendered inventory feed. This is a platform configuration change,not a full rebuild. The payoff: every AI shopping assistant can suddenly see and recommend this dealer's inventory.
2022 Mercedes-Benz GLC 300 4MATIC® SUV under 35000 miles
╔══════════════════════════════════════════════════════════╗
║ 911 — AUDIT CANNOT PROCEED
╚══════════════════════════════════════════════════════════╝
REASON:
The Pre-Owned inventory page at drivemb.com loaded, but the page
title reads "0 Pre-Owned Cars, Trucks, SUVs in Stock." The inventory
is JavaScript-rendered via Dealer Inspire's platform. This tool is a
crawler, not a browser — it reads raw HTML, not dynamically loaded
content. Zero vehicle records (VINs, prices, mileage, equipment)
were readable in the response.
WHAT THIS IS NOT:
This is not a statement about what human shoppers see.
Shoppers on a phone browser see inventory normally after the page
loads. This is a tool limitation, not a dealer website failure.
WHAT CANNOT BE PRODUCED:
— A real price comparison
— A real equipment analysis
— A real scorecard
Any output generated without confirmed inventory data is fiction.
This audit will not produce fiction.
WHAT THE DEALER CAN DO:
To enable this audit, ask your website vendor (Dealer Inspire) to
implement server-side rendering (SSR) or a static pre-rendered
inventory feed. This makes individual vehicle listings readable by
crawlers, AI assistants, and audit tools — without changing anything
human shoppers see.
WHAT TO DO NEXT:
Option A: Run this audit on a dealer whose inventory IS crawler-readable.
Option B: Provide a direct VDP URL for a specific 2022 GLC 300 4MATIC
listing. If that VDP loads with price, mileage, and equipment
in the raw HTML, the audit can proceed from there.
[SECONDARY FINDING — AI DISCOVERABILITY GAP]
The same JavaScript-rendering issue that blocked this audit alsoblocks Google's crawler, ChatGPT, Perplexity, and other AI shoppingassistants from reading Mercedes-Benz of Hagerstown's inventory. Agrowing share of luxury car shoppers start with an AI query — "find mea 2022 GLC 300 under 35k miles near Hagerstown" — and this dealer'svehicles are invisible to every one of those tools.
Severity: MEDIUM-HIGH
Recommended fix:Ask Dealer Inspire to enable server-side rendering (SSR) or a staticpre-rendered inventory feed. This is a platform configuration change,not a full rebuild. The payoff: every AI shopping assistant can suddenly see and recommend this dealer's inventory.
Used 2024 Subaru Ascent Onyx Edition For Sale in Carrollton, OH | #72852
2024 Subaru Ascent Onyx Edition - 37336 miles
Fetched: 2024 Subaru Ascent Onyx Edition
Fetched: 2024 Subaru Ascent Onyx Edition
Phase 2 — INVENTORY GATE: ✅ PASS
Real listing confirmed:
Searched the web
Searched the web
Good — I have enough confirmed Carvana data from search results to use as the closest comparable. The best Carvana match for a 2024 Ascent Onyx Edition nearest to 37k miles is the 29k-mile unit at $35,590 (from search result index 4). Now building the full audit.
Reading frontend design skill
Reading frontend design skill
Now I have everything I need. Building the full HTML audit file.
Created a file, read a file
Created a file, read a file
The audit is ready. Save the file as audit.html and open it in any browser.
Here's what the audit found at a glance:
Phase 2 — PASS. Real inventory confirmed: VIN 4S4WMAHDXR3405043, $33,100, 37,336 miles, Stock #72852.
The big finding: Huebner has a genuinely strong vehicle at a genuinely strong price — $2,490 below the closest Carvana Onyx Edition — but the mobile SRP is bleeding shoppers before they ever see that deal. Carvana wins the "find the car" test handily because their filter flow is built for a thumb, not a cursor.
Overall scores: Carvana 8/10 — Huebner 7/10. Not a blowout, but driven entirely by SRP friction. The VDP content at Huebner actually edges Carvana out because the listing is better documented.
The one move to make today: Stop sending buyers to the SRP. Every text, email, and social post should link directly to the VDP URL. One tap, straight to 43 photos and the full story. That alone eliminates the biggest friction point without touching the website.
2024 Subaru Ascent Onyx Edition - 37336 miles
Fetched: 2024 Subaru Ascent Onyx Edition
Fetched: 2024 Subaru Ascent Onyx Edition
Phase 2 — INVENTORY GATE: ✅ PASS
Real listing confirmed:
- VIN: 4S4WMAHDXR3405043
- Stock: #72852
- Price: $33,100
- Mileage: 37,336
- Trim: 2024 Subaru Ascent Onyx Edition, Magnetite Gray Metallic
Searched the web
Searched the web
Good — I have enough confirmed Carvana data from search results to use as the closest comparable. The best Carvana match for a 2024 Ascent Onyx Edition nearest to 37k miles is the 29k-mile unit at $35,590 (from search result index 4). Now building the full audit.
Reading frontend design skill
Reading frontend design skill
Now I have everything I need. Building the full HTML audit file.
Created a file, read a file
Created a file, read a file
The audit is ready. Save the file as audit.html and open it in any browser.
Here's what the audit found at a glance:
Phase 2 — PASS. Real inventory confirmed: VIN 4S4WMAHDXR3405043, $33,100, 37,336 miles, Stock #72852.
The big finding: Huebner has a genuinely strong vehicle at a genuinely strong price — $2,490 below the closest Carvana Onyx Edition — but the mobile SRP is bleeding shoppers before they ever see that deal. Carvana wins the "find the car" test handily because their filter flow is built for a thumb, not a cursor.
Overall scores: Carvana 8/10 — Huebner 7/10. Not a blowout, but driven entirely by SRP friction. The VDP content at Huebner actually edges Carvana out because the listing is better documented.
The one move to make today: Stop sending buyers to the SRP. Every text, email, and social post should link directly to the VDP URL. One tap, straight to 43 photos and the full story. That alone eliminates the biggest friction point without touching the website.
- Apr 7, 2009
- 5,041
- 2,462
- Awards
- 10
- First Name
- Joe
Hey Josh, this is pretty bad ass! DISC? Hell Yea!
DR Forum peeps, @jscole86 and I have been swaping best practices for a decade. Josh and the team at Huebner execute at a level that few get to. Wayyy back in the day, back in the Uncle Joe Diary 1.0 days, Josh and I built the most bad ass comment generators using php and tags and car guy logic.
much respect!
p.s. Josh, Let me know what mods you'd like (or your making
Big finding is we need to make our UX / UI much easier for customers on mobile, and unfortunately for most franchise sites, that's hard to do with the vendors we are able to use.
Now we just need to send a list of ALL of our VDP's to Claude Cowork to analyze with this prompt while we sleep, so we can wake up and fix any issues quicker! This was only the 1st and only vehicle I picked, and sure it seemed to do fine, but I can guarantee you I'd find plenty of others that need to be improved content wise.
- Apr 7, 2009
- 5,041
- 2,462
- Awards
- 10
- First Name
- Joe
EVERYONE NEEDS TO KNOW AI = Automotive Intelligence.
AI powered Car Shopping assistants will be all over your inventory. Your VDP has never been more important.
Equipment Analysis: The Huebner unit is loaded. Carvana's Onyx listings at similar mileage are priced $2,490 to$5,490 higher for fewer miles.
The equipment gap is real and the price gap is meaningful.
VALUE SUMMARY Huebner's $33,100 price on this 37k-mile Onyx Edition represents the strongest documented valuein the current market.
AI powered Car Shopping assistants will be all over your inventory. Your VDP has never been more important.
Equipment Analysis: The Huebner unit is loaded. Carvana's Onyx listings at similar mileage are priced $2,490 to$5,490 higher for fewer miles.
The equipment gap is real and the price gap is meaningful.
VALUE SUMMARY Huebner's $33,100 price on this 37k-mile Onyx Edition represents the strongest documented valuein the current market.
