2024 Stats from FRIKINtech are out: 2024 Automotive Dealer and Consumer Benchmarks Report | FRIKINtech

I'm looking forward to this thread! I was in a similar situation back in 2016. I attempted to create a "Dear Diary" right here on DealerRefresh. Don't give up on documenting Bill! It sounds like we are going to witness a serious turnaround.

My biggest advice - Don't worry about fancy marketing and driving traffic to the srote without the process in place to CONVERT (People + Process). This is something I overlooked.

Carsten said:

That's almost a good reply, but this gave it away:
"Thanks for laying out the legacy systems so clearly. I’d like to introduce you to what’s coming next ..."
Granted Ai has come a remarkably long way you can still tell.

Click to expand...

Dude, Prompt engineering sets up context.

You are seeing the me talking to AI, talking to you.​

​

Here is the prompt I created (see the context):

"You are a CTO preparing your team for the paradigm shift that AI will bring. I am the founder and I support your plan. Below is an internal forum thread by a dev named Carsten. Read the thread and craft a reply from your POV that introduces Carsten to the newest information with citations that will open his mind. For example, 'reasoning' itself is creating amazing results. Create the narrative from a CTO's position, speaking to a dev that sees only the weakness of LLMs."

AI's impact to daily life is going to be bigger than the internet itself. For example, AutoMagicLabs.ai has pivoted to become a LLM-centric platform where we feed the LLM data and craft prompts to get shit done.

IMO, this thread is going to be yet another DealerRefresh classic.

Lurking vendors, @BillVaughnAlWestNissan's play by play here reveals how deeply connected to the sales floor great GSMs' are. If you are a leader in the SAAS space, this is an EPIC USER STORY to share with your team.

SAAS leaders, be like Bill V, partner with Bill, roll up your sleeves and offer your SAAS at no charge. You will discover that lead gen. is such a small part of 'getting cars over the curb'.

BillVaughnAlWestNissan said:

You know that I'm 100% in on your product...

Click to expand...

Bar Rescue comes to Auto...​

​

Dude, you are going to nail it and I want to be a player in this turn around. Our platform is in the midst of a big upgrade, once we go live, I'll find a way to make this happen.

Carsten said:

I think the wold has evolved enough that people have a particular vehicle in mind when they make a call.

The interactions probably looks like this:

Samantha: Hi, this is Samantha, how can I help you today.
Caller: I'm calling about that 2015 F150 that I saw online.


How should Samantha respond?

Samantha: ...

Click to expand...

You're absolutely right that many shoppers have a specific vehicle in mind when they reach out, especially those further down the purchase funnel. But where AI can enhance this interaction is in Samantha’s response. Instead of just answering basic questions, AI-powered systems can instantly pull up real-time inventory details, suggest similar vehicles if that one has sold, and even estimate trade-in values based on the caller’s information.

For example, an AI-assisted Samantha might respond with:

‘Great choice! That 2015 F-150 is still available. It has X miles and comes with [notable feature]. Would you like to set up a test drive? Also, we have a 2017 model with lower miles and an upgraded package—would you like me to send you the details?’

AI ensures that Samantha isn't just answering a question but guiding the caller toward a potential sale while maximizing upsell opportunities. In today’s market, it’s not just about having the right vehicle—it’s about making sure the customer knows all their options and giving them a reason to take action right away.

BillVaughnAlWestNissan said:

38 of our 65 new cars are over 99 days old. 17 of those are over 148 days old. 4 of those are over 198 days old.

Click to expand...

Dude, I want to help! We're closing in on our new platform. If you're up to it, when I have some bandwidth open up, I'll find a way to plug you in and help you sell more cars.

BPFUNK said:

Thanks for the response...I'll have to dig up a contact at Credit Union of Colorado. We snuck our way into CUDL for a short period of time, but couldn't justify the monthly fee for our deal flow (It was like $750 a month). Many of our deals tend to be out of state, so a go-to national lender that has a polished customer experience what I need in a perfect world.

Click to expand...

Trina Larson
larsont@cuofco.org
(303)906-1368

BillVaughnAlWestNissan said:

Best of luck. @Tallcool1 is a PHENOMENAL source of information for a store like yours. Trust him.

Click to expand...

Thanks Bill. Yes, it used to be $350 as you remember. CUDL has done a really good job of slowly and steadily raising the fee. I believe mine is now $615. However, it is worth every penny. I have one lender that I use outside of CUDL and ironically it is another Credit Union that prefers RouteOne to CUDL...probably because the lender fees are so much lower. I am not sure.

alexgehrke said:

My business partners and I have been developing an AI calling agent that has access to live inventory of your dealership. This is just one part of our CRM that we are working on right now.

Here is a link with a demonstration of the AI caller: Login to view embedded media View: https://youtu.be/4a8dkytQzLc


I was just hoping to get some feedback from any dealerships currently using one or even just a dealership that is looking to get into AI automation. We want to make this as good as possible, so any feedback or comments are very appreciated!

Click to expand...

Conversational AI is evolving rapidly, and we believe it’s essential for dealerships to integrate AI agents into various aspects of daily operations.

Kudos on developing Samantha—creating an AI that can handle the many variables of a customer interaction is no small task. It’s incredibly challenging to anticipate every possible question, provide accurate and relevant responses, and still guide the conversation toward a meaningful outcome, like setting an appointment.

This is a strong effort, and I’m sure you’ll continue refining it.

Since you asked for feedback, here are a few areas where we believe there’s room for improvement—not as a critique, but as opportunities to enhance the natural flow and effectiveness of the conversation.



This conversation between Samantha, the AI assistant, and John, the potential customer, does feel unnatural and rushed. Here are the key issues and ways the AI assistant could have handled the conversation better:

Critique of the Conversation

1. John’s Openness Feels Unnatural
   • John starts by saying he's "thinking about buying a new car" without specifying any details. Most buyers already have at least a rough idea of what they’re looking for—whether it's an SUV, a truck, a sedan, or something else.
   • He then suddenly defines his needs in a very structured way: "I like to go camping, biking, and skiing." While this isn't impossible, it sounds scripted rather than how a real customer would gradually reveal their needs.
2. Samantha’s Inventory Response Lacks Depth
   • Samantha jumps from John’s general needs to listing vehicles without explaining why these cars are good for his activities.
   • She provides only the model, year, and mileage—but nothing about their features, trim levels, or why they might suit his adventures.
   • Instead of giving a brief description of each (e.g., "The Wrangler is a great off-road SUV, while the Santa Fe is a comfortable all-wheel-drive option"), she just lists them as if John is supposed to know all the details.
3. The Phrase "Which One Catches Your Eye?" Feels Odd
   • John hasn’t seen the vehicles. A better way to phrase this would be:
     "Would you like more details on any of these, such as features, history, or pricing?"
   • The lack of price details is also a problem. John specified a $60,000 budget, but Samantha doesn’t confirm how these vehicles fit within it.
4. John’s Quick Decision Feels Forced
   • He goes from hearing three basic options to saying, "Can I test drive the Mitsubishi?" without asking for any details like pricing, features, or condition.
   • In reality, customers usually ask follow-up questions before committing to a test drive.
5. No Discussion of Test Drive Preparation
   • Samantha should have confirmed if John wants any additional details before locking in the test drive.
   • She also missed an opportunity to set expectations:
     • Will he need to bring anything (like a license and insurance)?
     • Will a salesperson be there to answer questions?
     • Will the car be ready and gassed up for him?

How the AI Assistant Could Improve

1. More Natural Discovery Questions
   • Instead of asking, "What kind of car are you interested in?" Samantha could guide the conversation with more real-world questions:
     • "Are you looking for something rugged, like a Jeep, or something more comfortable for long trips?"
     • "Do you need all-wheel drive for winter skiing trips?"
   • This way, John can express preferences organically, rather than giving a structured "adventurer" description.
2. Present Vehicles in a More Engaging Way
   • Instead of a plain list, Samantha should provide:
     • A brief selling point for each car (e.g., "The Wrangler is a classic for off-road fun, while the Santa Fe has great cargo space for your gear.")
     • Basic pricing to confirm they fit within John’s budget.
     • Whether the vehicles are new or used (which wasn’t mentioned at all).
   • Example:

"I found three great options for you. There's a 2014 Jeep Wrangler Unlimited Sport, a rugged off-roader that's perfect for camping and skiing trips, priced at $XX,XXX. We also have a 2016 Hyundai Santa Fe GL with all-wheel drive, which is great for winter conditions, and a 2019 Mitsubishi ASX, which is the newest option with a smooth ride. Would you like more details on any of them?"

1. Give John More Control Over His Decision
   • Instead of rushing to "Which one catches your eye?" the AI should encourage more interaction:

"Would you like to know more about their features, pricing, or test drive availability?"

1. Make the Test Drive Setup More Customer-Focused
   • Instead of just jumping to a time slot, Samantha should:
     • Offer flexibility: "Would you like to come in this week to check them out and take a test drive?"
     • Set expectations: "You’ll be able to test drive and ask any questions with one of our specialists. Do you have any preferences for what you'd like to check during the test drive?"
2. Leave the Door Open for Further Engagement
   • At the end, instead of just saying "Is there anything else I can help you with?" Samantha could:
     • Offer a follow-up: "I'll send you the details on the Mitsubishi ASX, and you can reach out if you have any questions. Sound good?"
     • Remind him of next steps: "You'll get an email confirmation with your test drive appointment. If anything changes, feel free to reschedule using the link!"

Final Thoughts

• This AI conversation rushes John through the sales funnel unnaturally—from having no idea what he wants to booking a test drive within a minute.
• Samantha needs to be more conversational and customer-focused, guiding John without making assumptions.
• She should provide better details on vehicle features, pricing, and suitability.
• The phrasing needs to be adjusted to feel natural, especially with questions like "Which one catches your eye?" when John hasn’t seen anything.

Here’s a more natural and customer-focused rewrite of the conversation between Samantha (the AI assistant) and John (the potential customer). This version improves flow, engagement, and detail while guiding John without making assumptions or rushing him.

Improved Conversation Flow

Introduction

Samantha: Hi, thanks for calling Vita Motors! My name is Samantha. How can I help you today?

John: Hey, this is John. I'm thinking about buying a car. I saw you guys online.

Samantha: That’s great, John! Are you looking for something specific, or are you still exploring your options?

Understanding John's Needs

John: I’m not exactly sure yet, but I want something I can take on adventures. I like to go camping, biking, and skiing.

Samantha: That sounds like a lot of fun! So you’re looking for something that can handle different terrains and carry your gear? Do you usually drive off-road, or do you want something more comfortable for long trips?

John: A little of both. I’d like something that can handle rougher roads but still be comfortable on highways.

Samantha: Got it! Do you have a preference for SUVs, trucks, or crossovers, or are you open to different styles?

John: Probably an SUV or something similar.

Introducing Relevant Options

Samantha: That makes sense! I can check what we have in stock that fits your needs. Do you have a budget range in mind?

John: I’m looking to stay under $60,000.

Samantha: Perfect. I’ll pull up a few vehicles that fit your budget and are great for adventures like yours. One moment...

(brief pause for effect)

Samantha: Here are a few options that might work for you:

• 2014 Jeep Wrangler Unlimited Sport – A classic off-road SUV with 87,300 miles. It’s great for rugged trails and comes with removable doors and a hardtop.
• 2016 Hyundai Santa Fe GL – A spacious all-wheel-drive SUV with 80,900 miles, good for long trips and carrying gear.
• 2019 Mitsubishi ASX – A newer compact SUV with 103,727 miles. It has good fuel efficiency and a smooth ride.

Would you like more details on any of these, such as features, history, or pricing?

John Expresses Interest

John: The Mitsubishi sounds interesting. Can you tell me more about it?

Samantha: Absolutely! The 2019 Mitsubishi ASX is a fuel-efficient compact SUV with all-wheel drive, making it good for winter roads. It has a spacious cargo area for camping gear and a modern infotainment system. It’s priced at $XX,XXX.

Would you like to come by and check it out in person? We can also set up a test drive if you’d like.

Scheduling a Test Drive

John: Yeah, I’d like to test drive it.

Samantha: Great! We have availability for test drives this week. Do you prefer a weekday or weekend?

John: Tomorrow works.

Samantha: Perfect! We have test drive slots open at 8 AM, 11 AM, and 2:30 PM tomorrow. Which time works best for you?

John: Let’s do 2:30.

Samantha: You’re all set for a 2:30 PM test drive of the 2019 Mitsubishi ASX tomorrow. I’ll send you a confirmation email with the details. You can reschedule if needed using the link in the email.

Do you have any other questions, or would you like me to send details on any other vehicles before your visit?

John: No, that’s all for now. Thanks, Samantha!

Samantha: You’re welcome, John! Looking forward to seeing you tomorrow. Have a great day!

Appreciate any insight from those familiar with SEO.

How much do SEO keywords matter in a vehicle description / comments? Inventory comes and goes so frequently that I never really thought it made much of a difference, so we’ve always written them to be informative and engaging from the customer’s perspective, without trying to fit keywords in there.

BillVaughnAlWestNissan said:

Pricing issues... pricing issues everywhere!

3rd party sites and the website don't have prices that agree with one another. We have to be ensuring pricing accuracy regularly. Why do dealers just hope the pricing is accurate! Verify!

We have now done a complete retraining and both internet leads and phone up. Now time for the follow up training. I don't believe in training. I believe in training in the classroom, then training on the job, then following up on the training, and then holding people accountable to said training. Make sure they have the training, the tools they need to be successful, they know the expectation, and then hold them accountable to that expectation.

I'm taking a look at commissions earned previously and my expectations of them going forward. With the losses the dealership is taking, I don't believe salespeople are making anything off of what they are selling. With where we're at, I'm thinking we are going to need thin margins for a period to move metal, so I don't see it getting any better. I'm thinking this might be a good time to move to a non-commissioned sales staff and use it as an organic selling point.

Thoughts?

Click to expand...

It’s a perfect time for that pay plan switch. Did you have that pay structure at your previous store? That’s our structure and we made that change around the same reasons you listed. Happy to talk through it all with you and share ours.

What inventory tools does the store use? Vauto? Homenet?

I highly recommend VinCue.

BPFUNK said:

Any recommended 3rd party financing vendors? We are an extremely small independent dealer, but tend to traffic in higher priced vehicles $50k+. I don't mind treating financing as a customer service at the moment, as we just don't sell enough cars to put a ton of resources into F&I at the moment. We are in Colorado.

Click to expand...

You may want to look into a relationship with the Credit Union of Colorado. They are on CUDL which you may not want to invest in. However, you may be able to establish an actual relationship with a local branch where you physically send customers to them or maybe they have some type of a Fax App System in place.

Carsten said:

So, Ai is just a percentage relation to words that come before and after.

Click to expand...

Carsten,

Thanks for laying out the legacy systems so clearly. I’d like to introduce you to what’s coming next: AI that doesn’t just regurgitate old code or static patterns but actually “thinks” through problems.

Recent breakthroughs in large language models (LLMs) have moved beyond simple next-word prediction. Using techniques like chain‑of‑thought prompting, new models are now able to decompose complex problems into a series of intermediate reasoning steps before providing a final answer. In practical terms, models such as OpenAI’s o1 (internally called Strawberry) and DeepSeek’s R1 now solve challenging math, coding, and scientific tasks at levels comparable to human experts.

These models spend extra “thinking time” during inference—much like a chess player considering several moves ahead—allowing them to refine their strategies and even self-correct before answering. For example, on rigorous tests like the International Mathematics Olympiad qualifying exam, o1 has dramatically outperformed previous models. This isn’t just an incremental update; it represents a shift from “dumb” output generation to a dynamic, human-like reasoning process.

I know your expertise is rooted in the tangible aspects of our systems—microservices, APIs, and the cloud—but imagine integrating these smarter models into our architecture. They could help automate debugging, optimize code, and even assist in strategic planning by analyzing vast datasets in ways we never could manually.

I invite you to explore these advancements with me and consider how we might pilot these models in our projects to drive our next-generation innovations.

Below are several key references that outline these developments, along with summaries and their URLs:

1. Chain-of-Thought Prompting Elicits Reasoning in Large Language Models
   This seminal paper by Jason Wei et al. introduces the chain-of-thought technique, showing how LLMs can be prompted to break down complex problems into step-by-step reasoning. This method has been crucial in unlocking improved performance on challenging tasks.
   URL: Chain-of-Thought Prompting Elicits Reasoning in Large Language Models
2. OpenAI Launches New Series of AI Models with ‘Reasoning’ Abilities (Reuters)
   This Reuters article details OpenAI’s recent release of the o1 model (code-named Strawberry), which employs chain-of-thought reasoning to tackle complex problems in science, coding, and math—demonstrating a significant leap over previous models.
   URL: https://www.reuters.com/technology/...ies-ai-models-solve-hard-problems-2024-09-12/
3. OpenAI Announces a New AI Model, Code-Named Strawberry, That Solves Difficult Problems Step by Step (Wired)
   Wired’s coverage explains how the new o1 model reasons through problems step-by-step—“thinking aloud” before arriving at a final answer. It highlights the model’s enhanced performance on advanced tasks and its potential impact on our industry.
   URL: OpenAI Announces a New AI Model, Code-Named Strawberry, That Solves Difficult Problems Step by Step
4. OpenAI’s o1 Model is Inching Closer to Humanlike Intelligence – But Don’t Get Carried Away (Business Insider)
   This Business Insider article discusses how o1’s extended reasoning time allows it to achieve results that resemble human problem-solving, particularly in STEM fields, while noting that challenges like errors and hallucinations still remain.
   URL: OpenAI's o1 model is inching closer to humanlike intelligence — but don't get carried away
5. What It Means That New AIs Can “Reason” (Vox)
   Vox provides insights into the significance of AI models that “think” before answering. It describes the internal chain-of-thought process that enhances the accuracy and robustness of outputs, as well as the dual-use risks associated with these advances.
   URL: What it means that new AIs can “reason”

Looking forward to your thoughts on how we can integrate these exciting advancements into our next steps.

Best regards,
GPT o3-mini-high

BillVaughnAlWestNissan said:

Yesterday we started the process of eliminating all vendors. No, seriously. EVERYTHING. All 3rd party sites, Nissan Buy At Home, Nissan 3rd party leads, SEM, all of it. I want to know what Al West Nissan looks like completely naked so I know how much every piece we add back in the future moves the needle.

Click to expand...

This is completely insane and I absolutely love it.

In the car business, it is anything that came out in the past 10 years. Unfortunately, too many systems are handcuffed to languages that are many decades old.

As for AI, my post stated that newer coding languages can better work with it. It isn't that something cannot be done with older stuff. It is easier to utilize an AI with newer tech.

kylewilliam said:

Recently, an insurance company canceled my car insurance due to an error on a previous policy where I was a named driver, not the policyholder. The policy incorrectly listed two historical accidents as fault accidents, when they were non-fault. Although this has since been corrected, I now have to disclose on any new proposals that my insurance was refused or canceled. As a result, I'm noticing that my quotes for my other vehicles are higher. How much of this increase is due to this incident, and is there any compensation I can seek?

Click to expand...

Sucks to hear that. Do you think this has to do with AI? <trying to get this back on topic>

This is waaaaaaay out there. Why would a dealer want to do this? Where is this car going to be driven? How much money is the dealer going to make?

Ya got some holes in this ask.

Carsten said:

If I knew what you were aiming for, it's easier to reply so please consider that when I am writing all this.

Click to expand...

There are many ways to answer this and you did an excellent job. One of the reasons our government's systems are so poor is because they never updated these systems. One part of an update means utilizing newer programming languages. In the case of our federal government, people who knew how to code in those old languages retired. The newer employees were told to just leave it alone and they'd contract on old guy to come do something if things got bad. Well, those old guys are dying. Good strategy!

There are also the limitations of some older things like PHP. And the old school ways of tabling data. Newer codes have better rendering, ways to be more precise with an object model, and can work with the latest stuff easier (like AI). The downside comes from trying to integrate with older systems. You may have to blend some stuff that takes some workarounds. Also, newer languages can get frequent updates, but those updates can come with broken parts.

To your point Carsten, those updates eventually sort things out.

Richard_Lloyd said:

A Nissan Kicks has massaging seats?!

Click to expand...

Ooooo... You know your cars! mad respect from one car guy to another.