Three Ways AI Is Already Changing Your Dealership
Episode Focus
Here's a split version — I've alternated so each segment has a natural back-and-forth rather than one person monologuing. Swap any pairing on the fly if it feels more natural live.
Cold open
-
You: Set up the three-shift frame — AI talking to leads, AI shopping on the customer's behalf, AI in the financing decision.
-
Co-host: Land the line — "not one big AI trend, three different things happening to the car deal right now." Tee up the episode.
Segment 1 — AI inside the dealership
-
You: Drop the stat — 57% of dealership staff already use AI (70% among execs/DPs). Reframe: not "should we use AI" but "what should we let it do."
-
Co-host: Riff the task list (lead response, scheduling, follow-up, descriptions, reporting, marketing, upsell), then land the core idea — AI's real value is freeing up salesperson time. Ask what they should do with it (calls, showroom, relationships).
Segment 2 — AI as the shopper
-
Co-host: Flip the lens — customers use AI too. Cite Fullpath: LLM-driven site traffic climbing hard (43% MoM jump last Sept, now 15x YoY).
-
You: Walk the new funnel — AI → recommendation → website → lead. Riff on what AI "reads" about a dealer (inventory, pricing, reviews, FAQs, descriptions). Land the line: it's not "can Google find me," it's "can AI understand and recommend me."
Segment 3 — Both sides have AI
-
You: Paint the scene — customer's AI and dealer's AI interacting before a human's involved.
-
Co-host: Riff on what's left for the salesperson (fit, deal structure, trade-ins, the "weird thing"). Land core idea: easier AI makes simple stuff easy, humans get more valuable for the complicated stuff.
Segment 4 — AI and lending
-
You: Explain AI underwriting looks at far more variables than traditional models. Read the fair lending disclaimer here (or point to show notes).
-
Co-host: Drop Upstart's numbers — 41% more approvals at 33% lower rates vs. a comparable traditional model, 91% of loans fully automated. Reframe as dealer questions: speed, throughput, reaching buyers a traditional model misses. Optional scale color: Q2 2026 originations $4.2B, +50% YoY.
Segment 5 — What happens to humans
-
Co-host: Run the contrast pairs back-and-forth (speed/judgment, information/context, scale/relationship, repetitive/weird deals, 24/7/gut check) — good for quick volley.
-
You: Counter the "employees resist AI" myth — 90% say they're comfortable learning new tech. Land core idea: automate the task, not the relationship.
Segment 6 — The AI audit (closer)
-
You: Give the three-question takeaway — what's it doing, where's the data from, where does a human take over.
-
Co-host: Riff the "always needs a human" list (upset customer, complex financing, weird trade, conflicting info, big opportunity, AI's blind spots). Close together on: winning isn't about having the most AI, it's knowing where to use it.
Cold Open
Your next customer might talk to AI before they ever talk to your salesperson.
Your dealership might have AI responding before your salesperson even sees the lead.
And when that customer applies for financing, AI could be involved there, too.
This isn't one big AI trend coming someday.
It's three different changes happening to the car deal right now.
So today we're breaking down where AI is showing up, what it's actually doing and where the human still matters.
Segment 1: AI Is Already Working Inside Dealerships
A 2026 Reynolds and Reynolds survey of more than 500 dealership employees found 57% already use AI in some capacity as part of their job.
Among executives and dealer principals, usage was even higher at 70%.
So the conversation has moved past: “Should dealerships use AI?”
The better question is: “What should we actually let it do?”
Where It's Showing Up
AI tools are already helping with things like:
-
Lead responses
-
Appointment scheduling
-
Customer follow-up
-
Vehicle descriptions
-
Reporting and data analysis
-
Marketing
-
Upsell recommendations
Dealer Lens
The value isn't:
AI can write an email.
It's:
What are your people spending time doing today that they shouldn't still be doing manually?
If AI takes 10 repetitive tasks off a salesperson's plate, what should that salesperson be doing with the time they get back?
Calling customers?
Working the showroom?
Following up on opportunities?
Building relationships?
Core Idea: The best use of AI may not be replacing the salesperson. It's giving the salesperson more time to sell.
Segment 2: AI Is Becoming the Shopper
Now flip the perspective.
The dealership isn't the only one using AI.
The customer is, too.
Fullpath's 2026 Auto Intelligence Index found AI and large language models emerging as a meaningful source of dealership website traffic.
Visits to dealership websites originating from LLMs increased throughout 2025, including a 43% month-over-month jump in September.
more recent figure from fullpath is 15X YOY
That means customers are increasingly doing something like:
“Find me a three-row SUV under $45,000 with good reliability and dealerships near me that have one.”
And AI can narrow the field before the customer ever searches your inventory directly.
Dealer Lens
That's a major shift.
For years, dealers optimized for:
Google → website → lead
Now you potentially have:
AI → recommendation → website → lead
So what information is the AI using to decide whether your dealership gets recommended?
Data Points to Riff From
AI systems may encounter information about:
-
Inventory
-
Pricing
-
Reviews
-
Website content
-
Business information
-
Vehicle descriptions
-
FAQs
-
Other publicly accessible information about the dealership
That creates a new version of the old SEO conversation.
It isn't only:
“Can Google find me?”
It's increasingly:
“Can AI understand me well enough to recommend me?”
Dealer Opportunity
This makes fundamentals more important, not less:
Clean inventory information.
Accurate pricing.
Useful vehicle descriptions.
Consistent dealership information.
Strong reviews.
Clear answers to common customer questions.
Core Idea: Your next shopper may not search for your dealership. AI may choose it for them.
Segment 3: What Happens When Both Sides Have AI?
I'd add this because it connects Agents and Virtual Shoppers instead of treating them as unrelated topics.
Imagine the transaction:
Customer's AI:
Find me the right vehicle.
Dealer's AI:
Respond to the customer.
Now an AI-assisted shopper is potentially interacting with an AI-assisted dealership before a salesperson ever enters the conversation.
Dealer Lens
If AI can answer:
What's the price?
Is it available?
What's the range?
Can I schedule a test drive?
then the salesperson's value shifts toward:
Which car actually works for you?
How do we structure this deal?
What happens with your trade?
How do we solve the weird thing that doesn't fit neatly into the system?
Core Idea: The easier AI makes the simple questions, the more valuable humans become for the complicated ones.
Segment 4: AI Is Changing the Lending Decision
Traditional underwriting has historically relied heavily on a relatively limited set of borrower information.
AI-based lending models can analyze substantially more variables and relationships between them when evaluating credit risk.
Upstart says its models currently consider a wide range of variables and engineered features.
What Upstart Reports
As of its April 2026 comparison, Upstart reports that its model produced:
41% more approvals
at
33% lower rates
compared with a hypothetical traditional model at the same loss rates/approval comparison methodology.
And 91% of Upstart-powered loans are fully automated, meaning they can move through the defined origination process without human intervention by Upstart.
For the dealer, that means the AI conversation becomes:
Can I get an answer faster?
Can I serve another customer?
Can I find qualified buyers a traditional approach might miss?
Can I get the customer further into the transaction before friction kills the deal?
Upstart Auto Context
Upstart reported strong growth across auto in Q1 2026 as part of broader platform growth, while total originations across the company increased 61%
Second Quarter 2026 Highlights
-
Originations: $4.2 billion, up 50% year-over-year (“YoY”). 558,014 loans originated, up 50% YoY.
-
Total Revenue: $365 million, up 42% YoY. Revenue from fees was $348 million, up 45% YoY.
-
Income from Operations: $14.6 million, compared to $4.5 million in Q2 2025.
-
Net Income: $16.5 million, up 195% YoY from $5.6 million in Q2 2025. Diluted net income per share was $0.16 compared with $0.05 in Q2 2025.
-
Contribution Profit: All-time high of $193 million, up 37% YoY. Contribution Margin was 55%, versus 58% in Q2 2025.
Core Idea: In lending, AI isn't just making the process faster. It's changing what information can be used to decide whether the deal works.
We probably need to include a disclaimer somewhere on the screen or show notes so we're not at fair lending risk
Segment 5: So... What Happens to the Humans?
AI can respond to customers.
Customers can shop with AI.
AI can help make lending decisions.
So yes:
What exactly are we supposed to do?
The same Reynolds survey that found 57% of dealership personnel already using AI also found 90% said they were comfortable learning and integrating new technology.
That's an interesting counterpoint to the idea that dealership employees are resisting AI.
Human Advantage
AI: Speed
Human: Judgment
AI: Information
Human: Context
AI: Follow-up at scale
Human: Relationship
AI: Repetitive work
Human: Weird deals
AI: Available 24/7
Human: Knows when something doesn't feel right
Dealer Lens
The question for management isn't:
“How many people can AI replace?”
A more productive first question may be:
“What should my best people never have to waste time doing again?”
Core Idea: Automate the task. Don't automate the relationship.
Segment 6: The AI Audit - How to be found (AI)
Since this is the finale, I'd finish with something dealers can actually take back to the store.
Look at every AI tool you're using and ask three things:
1. What Is It Actually Doing?
2. Where Does the Data Come From?
3. Where Does the Human Take Over?
Every AI workflow should have a point where someone has these answers.
What needs needs a person no matter what:
-
An upset customer
-
Complex financing
-
An unusual trade
-
Conflicting information
-
A high-value opportunity
-
Something the AI simply doesn't know