00:25
Hey everybody, welcome to another edition of The Daily Dealer Live!
00:29
I'm your host Sam Delerkin. Back with me co-hosting today is the Uli Di Martino. Uli, welcome back.
00:36
Thank you. Uli, guess where I am today? I'm in a secret location, but I'm gonna tell you exactly where I am.
00:44
Fort Wayne, Indiana. Tonight we've got the open house and the new facility celebration for our
00:51
Ziggler Subaru Fort Wayne facility. So as I said on Friday, if you're in the Fort Wayne market,
00:57
if you're around, come by, say hi. We've got a big party here between 5 and 7 p.m. Everybody's welcome.
01:02
Dealers, vendors, everybody across the Fort Wayne market. This is such a cool city and such a great
01:10
town for automotive. It's our second store here. We have a JD byrider store and the Subaru store is
01:16
second. We've had it what year and a half, almost two years I guess, but ground up Renault. That was
01:21
painful. It's painful to build stuff, Uli. I bet. No, it absolutely is. Well, congratulations to
01:27
getting it done and the grand reopening. Yeah, congrats to the teams for bringing it back. So
01:33
and for everybody joining our live stream today as a reminder, we're live across all CDG social
01:37
media platforms. Post those comments. We'll bring them into today's show and Uli, I feel like I say
01:43
this every time, but what a show we've got today. We've got a huge show coming up and first off,
01:48
just to tease a little bit about what we have coming up. Make sure you have your comments in
01:52
there because they help us with the show. Many have asked chat GPT, what do I do with it? What
01:57
do I do with AI? How do I engage and interact with it? And in fact, here at Ziggler, today we had a
02:02
team training call with, have you heard the name Jim Craig? Do you remember who Jim Craig is?
02:07
He was, he's an Olympic hockey gold medalist. He was the goalie on the famed Olympic team that
02:15
beat the Russians back in the 1980s. The Disney film Miracle on Ice is based on it. He does some
02:21
consulting with us. He comes in once a month and gets our team that wants to be leaders onto a call
02:27
and he asked today, he said, look, I'm old. Most of you are young. He said, one of the things with AI,
02:33
chat GPT and others, is he's like, who's going to explain it? Who's going to own it? And who's
02:38
going to explain it to somebody like me? And it really got me thinking, you know, one of the biggest
02:42
obstacles to the newest, greatest tech is figuring out how to connect it to other generations so
02:48
that it makes sense and so that it actually works. And today, for all the chat GPT talk we've had,
02:55
Yossi and the team at CDG put out an email, five hacks. Well, we've got an individual on today's
03:01
show, Killway, and his whole purpose in talking with us today is to explain some of the hacks he's
03:06
created in chat GPT to help run automotive better. So I'm excited to get to that. Plus,
03:13
we've got Auto Genius on Untapped Markets. But first, let's dive into today's headlines.
03:20
So first up today, August was a blockbuster month for new EVs. It's no shocker as the,
03:27
as the incentives tick down to zero. Sales hit a record 146,000 units up 14% from July
03:35
and nearly 18% higher than last year, lifting market share to an all-time high of 9.9% per
03:42
Cox Automotive. A whopping 13 brands notched their best EV sales this year with Tesla still
03:48
leading the pack at 55,000 units. Honda with only the prologue on the lot surged nearly
03:55
48% month over month to grab 6.4% market share. And Audi was the standout luxury brand up 65%
04:02
more than 7,300 units. However, that demand chewed through inventory. EV days supply fell
04:08
to just 62 days in August. By the way, that is intentional. Speaking as a retailer, looking
04:12
to your EVs before the incentives go away, it's the lowest level today on lots of 2025 down from
04:19
a June peak of 133. Toyota had the leanest supply of just 14 days while GMC sat on the highest at
04:27
139. What's the bottom line here? Well, buyers are clearly pulling ahead of the September 30
04:32
tax credit cutoff. Demand is strong now, but the real test comes in Q4. And you are seeing some
04:38
really interesting, like promotional stuff going on out there in retail right now. Tesla,
04:43
you pointed out in our prep call today, is advertising all over. Hey, just order that EV
04:48
by September 30 and you're okay. We all know or at least have been explained. You got to validate
04:54
this on your own that as long as the paperwork is signed, those EV sales are good, even if the
04:59
delivery ends up happening after the 30th, you got to test that. Don't take our word for it,
05:03
because I'm not legal advice, but that's what I've heard what I understand. So it's going to be a
05:08
fascinating race to the end. It's our last full week in retail sales, selling EVs under these
05:16
rebates. Next week, we've got a couple days left and then it's over. It's over and then we'll see
05:21
what the real demand is. So, all right, time for an M&A tracker, cdgbicell.com tracker, QR. There
05:31
we go. Doesn't that just make you happy? Every time, every time, what was the movie that says
05:37
something about angels wings and like on a Christmas tree and all that? Anytime a
05:42
car dealership is sold in the US, cdgbicell.com will report on it. That's right and we'll be there.
05:50
The highly automotive group is expanding its footprint again this time by picking up its
05:55
first Chevrolet store, the family owned group acquired Lakeside Chevrolet in Rockwell, Texas
06:00
from the Grady family and will rename the store Hilly Chevrolet according to the
06:05
deal advisor performance brokerage services. The move gives Hilly 11 dealerships across the
06:11
Dallas Fort Worth area, a great marketplace by the way, plus three more in Huntsville, Alabama.
06:17
I wanted to go to Huntsville, Alabama as a kid because I wanted to go to space camp. You probably
06:21
don't remember this. You're too young for it, but I remember space camp. The movie was in Huntsville.
06:27
I had the brochures I put down. I wanted to put a deposit. I wanted to be a space. I wanted to go
06:32
to space. So what's the big picture with these acquisitions? Props to them for picking up three
06:37
more in Alabama, Huntsville. With the addition of a domestic brand, Hilly could be opening the door
06:42
to a whole new customer base. And that is correct. And if you want to learn more about
06:50
acquisitions right now, go to cdgbicell.com. Next up today, Hyundai is recalling nearly
06:57
600,000 vehicles across three separate issues according to federal regulators. The biggest
07:03
recall is more than 568,000 Palisade SUVs from 2020 to 2025 for faulty seatbelt buckles
07:11
that may or may not latch. The company is also on notice for 31,000 Ionic 6 EVs because the
07:18
charging port doors could detach as well as a small batch of eight Ionic 5s over loose wheelbolts.
07:24
Hyundai says the Palisade defect likely affects only about 1% of vehicles, but owners will be
07:31
notified by November 10th and can get all repairs done free at dealerships, of course.
07:36
And for dealers, these recalls are of course golden opportunities in the service center,
07:40
potentially bringing thousands of vehicles back into service bays and giving stores
07:44
fresh customer touch points if those are handled correctly. And hopefully they are
07:50
at all dealerships everywhere across these United States. And finally up today,
07:54
last on the docket is Porsche, which just took a massive 1.8 billion euro hit after delays in its
08:01
EV rollout forced the automaker to slash its 25 outlook. The company now expects an operating
08:07
return on sales of just 2% down from its earlier forecast of 5 to 7. Parent company Volkswagen
08:14
is also lowering its guidance and recording a 3 billion euro impairment tied to Porsche.
08:20
Zooming out, this is Porsche's fourth guidance cut this year as cooling EV demand,
08:25
tariffs and high cost continue to drag on profitability. And as a result, the flagship SUV
08:31
that was supposed to launch as fully electric will now come in combustion and hybrid versions.
08:37
Good play, Porsche. I think hybrid is what customers want anyway. Take away here, Porsche's
08:42
retrenchment is the clearest signal yet of a major reset in electrification strategy
08:47
among European lawmakers. And with that, that's a wrap on today's news this Monday,
08:54
September 22nd. Excited for the show, man. I am too. All right, Uli. But before we get there,
09:01
let's talk guests. So if you're a dealer, if you have an opinion, if you want to be part of
09:07
this voice that is daily dealer live, part of this show, this is the voice of automotive.
09:13
There's nothing else like it Monday, Wednesday, Friday, 1pm Eastern. We're live here every week.
09:18
Dealers, if you want to join the show, go to cdgguest.com. Fill out our intake form to be
09:23
considered for a future guest spot. We'd love to have you on the show. And then we can't go into
09:30
today's show without talking about our NADA party. The car dealership guy is back with our
09:35
second NADA party happening in Las Vegas on Thursday, February 5th. It's the hottest ticket
09:41
at NADA 2026 with special guests and top automotive personalities will be there as well.
09:47
We're not top automotive personalities, but we will be there. We refuse to not be there.
09:52
To be considered for a formal invite, just hit the link in the show notes,
09:55
request to join and fill out the questionnaire. So if you're listening to this not live,
10:01
go to the show notes. If you are live, you can follow the link in the comments down below. We
10:07
hope to see you all there. All right. Without further ado, let's turn to VP of Digital Ops at
10:12
Casa Auto Group, Jared Killaway. Jared, welcome to the show. Hey, good morning. Good morning,
10:18
Casa, everybody. It's great to have you, Jared. Welcome to the show. All right. Before we go in
10:23
and ask you how's biz, because we're going to ask that along with that, just tell us a little bit
10:28
about yourself and what you do at Casa Auto Group. Yeah. So I am the vice president of digital
10:34
operations over here at Casa. So I'm tasked with really working with the team on process,
10:39
management, execution of anything and all things digital. So been in the industry 22 years,
10:46
worked with a couple of notable people who have been on your show before.
10:50
But it's a lot of fun and I'm excited with the change in landscape and what we deal with day
10:55
in and day out. All right. Which notable people have you worked with? Let's name drop. The one
11:02
and only Mr. Brian Kramer. Oh, okay. Very good. You know what? Isn't he insightful? You get him
11:09
in front of you and just ask him all about all things used cars and he always has incredible
11:14
insights. I can tell you quite a bit of stories on that one, but yes, working together down at
11:19
Maine and putting together the Metaverse project and many of other tasks that he was pushing my way.
11:26
So it's a lot of fun to work with people like him and the team over at Casa that are really trying
11:32
to test the limitations of what we do with day in and day out. Yeah. So speaking of testing
11:38
limitations. So how's Biz? What's business look like today at Casa Auto Group? Biz is growing.
11:45
Yeah, we grew. We actually entered the Arizona market here earlier this year. So
11:51
dealing with some new stores, some new OEMs and continuing learning, but also taking in a new
11:57
demographic and new geolocation. It's really exciting for me to be able to have Arizona now on
12:04
the map and be able to test some of these things we're talking about today to really assess the
12:09
market of what it is there at that store. So speaking of that, that's a great segue. AI,
12:18
ChatGPT, it's just all the rage. It's a buzzword almost right now. And I'll be in rooms with
12:24
general managers, dealers and other people and everyone knows it's here. Everyone has had that
12:30
moment of aha when you see ChatGPT do something astounding. But what's tough is to connect that
12:36
aha of, I can't believe it can do that to execution. And as you know, Jared, we're all in
12:43
automotive 100% about execution. So tell us a little bit about, you created some custom GPTs.
12:50
Hey, at first, tell us what the heck that is. And then what led you to develop this?
12:56
Yeah. So in essence, what I did, I was able to aggregate all of our DMS data, CRM data,
13:04
inventory data, match that to our operational data of Google AdWords, Google Analytics,
13:13
website experience, social media channels, paid social results through Meta Business Manager.
13:20
In essence, was able to connect everything into one instance. So it is completely secure in private,
13:27
which I think we need to kind of start there. Is that privacy element of the data?
13:34
Everybody on the call, don't go take your financial statement or extract all your
13:39
customers and go and run it into ChatGPT. Why? Tell us why, Jared. Why is that a bad idea?
13:47
Because there's smart people like myself and Brian Craver and Ben Hadley that could go and extract
13:54
that through queries because that is open and public at that point.
13:59
So how did you isolate it? You used a word. You said it's in a separate environment. It's in a
14:06
protected environment. How do you do that? What does that mean?
14:11
So in essence, connecting through API integrations, everybody in the space has heard of API,
14:17
whether it's one of your CRM vendors, but ChatGPT has an API element which you can tie into.
14:25
So I have created a safe space that houses all of this data as one big pile
14:32
and then using that through the API with GPT to query that information that already exists.
14:39
So that is not public. And for these ChatGPT solutions that you've created,
14:45
you're about to share with us. What is the API that you use to create that safe,
14:51
protected space to upload this to? So that is, I built a data clean room,
14:59
so it's extracting everything and it's on AWS. That is something I can query at any time.
15:05
So it's not out of the box. It's not something I just went and bought off the shelf. It did take
15:10
some time to build it. So we had Steve, I think it was Steve Greenfield came on the show a few
15:16
weeks ago and he said, hey, in our world today, it's not the tool or the solution because the
15:21
bar to creating new AI solutions is low. It's inexpensive and he kept saying, hey, it's easy
15:29
to create the solution. He said it's the problem and he said dealers who are super sharp will
15:35
identify the problem and then they can actually quickly go to executing and creating a solution.
15:39
How did you get the skill set to be able to say, hey, I'm going to create an API. I'm going to
15:44
create this protected area. What's your background that helped you to get to that point?
15:50
So my background, I've been 22 years in automotive, both retail, OEM, vendor space.
15:57
My previous regime too, coming to CASA, I was actually with an AI company in automotive space
16:05
in the very early ages that has now executed. But I've learned a lot through the years and
16:11
it's one of the things of finding solution for things that you don't know are broken yet.
16:16
So it's trying to think five years in advance beforehand. So the one thing I used to get,
16:22
and I still get, it's the, hey, I'm going to an NADA 20 group. What's my average RY per
16:29
marketing dollar spend? Or, hey, I need to go sell 15 more centuries between now and the end of the
16:35
month. What do we do? And it gives the ability, okay, go look at all of our specials off our
16:41
website. Go to look at all of our historical sold data based upon zip code of, hey, historically,
16:47
this is where we sell into. This is the market that created that sale. This is their APR average.
16:54
This is the income level. This is all of these dynamics and then stack all of that together
17:00
to launch and deploy a specific target on that one model to that audience that we know has the
17:07
highest propensity to purchase. So it's just trying to think more quickly. I used to do it with
17:13
Excel spreadsheets, which used to take me an hour or two hours to do to be able to really configure
17:18
everything dialed in. It's now, I can get results in a matter of a minute. So the time savings.
17:24
So Yoga Cars had a great comment. Yoga Cars says, unless you anonymize your data off AI,
17:31
you're at the mercy of AI until the next data leak. Agreed?
17:36
Absolutely. Yeah, privacy. All right. So let's back up a little bit. You gave us a little bit
17:42
of a tease just a minute ago on this. What is the, so when you think about the different tools
17:47
you've created, when you think about the most successful tool, what is the problem that you
17:52
are looking to solve with this tool? The main problem really looking to solve is the ability
17:59
for the general managers to take action and take action quickly without having to necessarily call
18:05
me or send me an email or jump on a zoom call. It's the ability to give them, in essence,
18:13
jittered in a box that can answer a question if they're at a 20 group or NADA Academy that
18:19
need to get specific information. They can have it very, very quickly. And so, and so what,
18:25
give us an example of a question that they would come up with at a 20 group or in making
18:30
an ad buying decision. What, you know, what's our website bounce rate? What's our unique visitors
18:35
based upon sessions? What is our auto, what is our referral traffic coming from our auto genius
18:42
auto group site, etc. They may not have that knowledge right away, but they can pull it
18:48
and it'll give them everything in a digestible format. So before you talk about the tool you
18:54
implemented to give them to that point, I'm a GM at your auto group. I have this question about
18:59
bounce rate. I'm at a 20 group. Where do I go to get that? Do I log into a chat GPT? Do I log
19:04
into a tool? Where do I go? No, so you as GM have a secure link to your store and your store's GPT
19:11
data that you'll be able to query. Okay. And it looks like the chat GPT interface, right? But it's
19:18
secure and I have my own login. Okay. Yeah, it's similar to it. And then what I've done is built the
19:23
most frequent prompts. So kind of like the most frequent questions that we're getting asked to
19:29
it so they can go, you could literally say, okay, is it in this list of five or six things?
19:33
And you'd be able to click a button and it'll give you results right there. And is this a real
19:37
time integration like back and forth? Yeah, it'll call the safe space API cloud where the data is
19:45
housed. And then we're getting the information from Techion or DMS or whoever, whoever the provider is,
19:53
we're getting that information in real time to that safe space. All right, let's test this.
19:58
Do you also have an enterprise version that allows it to all flow up to multiple rooftops?
20:03
Are you able to like say, hey, for all of CAS Auto Group, here is the metric of the result? Yes?
20:10
So that is something that I'm actually working on right now. I had to get it on the store level,
20:15
child level first, and now it's going to roll up reports, correct?
20:19
Okay. All right. So let's test this live. You don't have to tell us which store, pick a store,
20:23
any store. Who's your number one most effective digital lead provider, aside from the dealership
20:31
website? Now, by the way, yeah, except for that, by the way, if you can do this, how cool is that
20:43
to actually be able to make decisions and advertising digital dollar spend that are effective
20:49
rather than having to scramble around and prep for a meeting beforehand. So let's
20:55
car now, car now. Interesting. Interesting. Okay. So it's not website, but it sits on site,
21:04
but that's the one by how much?
21:10
And, and Jared, does it ever make mistakes? Like, do you have to go back and say, hey,
21:15
wait, that's not right? Or is it pretty accurate?
21:18
In the beginning when it was training, it was, okay, I have to inspect the expect every single
21:23
time. And then I got into, okay, I can do five queries, six queries, and then I can go back,
21:30
something looked off, go and retrain. But now it's been, it's been running about 30, 45 days
21:36
consistently. Now, so what were I expected every day just to kind of recap, but you'll know if it
21:42
kind of short circuits, and yeah, just spazes out on you, and you can get a good sense of, oh,
21:48
yeah, your, your bounce rate is 90%. Okay, something's not working correctly. So you'll be able to
21:55
pick up on instances very quickly. So you're 30, your percentage that you're looking for is 4%.
22:03
Wow, that's good. That's good. And at the point you could get that for the entire group,
22:08
how powerful that's going to be. All right, so, so before you give us how the sausage is made in
22:14
between, you've been on only 30, 45 days, right? Are you able to quantify the impact of easily
22:21
having this information accessible to GMs to increase sales, increased, you know, better
22:27
spend on ads? I mean, like, is there a quantifiable impact of having this at this point?
22:34
Not yet. Not yet. I think it's going to be something that I'm going to like review here,
22:39
end of Q4, once I have two full quarters of data running through it to see what really
22:46
is the impact. But then also, how is it, how has it made everybody else's job easier? And what can
22:52
we learn from this information? But, and then you improve it, you make it better, right? Oh, well,
22:58
it then it comes down to, you know, we think of all the different tools we have, connecting it into
23:04
an equisrade or something like that from vehicle information. I don't have it connected yet, but
23:10
you know, what if we did that and then built out something that could go
23:15
scale and review every car for sale by owner on Facebook Marketplace and automatically pull the
23:23
ones to it and reach, I mean, it's endless of where it could go. Yeah, but it's right now. And
23:29
that's the exciting thing for me. It's okay. Here's where it is an infancy stage. Where can this
23:34
become in two, three, five years from now for impasse? So did you have to, did you have to talk
23:41
someone into allowing you to do this? Like the executive leadership team, when you went to them
23:47
and said, Hey, I got a great idea. I'm going to load all of our info into this AI tool, and it's
23:51
going to spit out automatically information to GMs. They're going to use it as a resource where
23:55
they like, whoa, wait, hold on, like privacy implications, data security, like all the things,
24:01
incorrect info. Was there somebody you had to sell quote unquote on the idea?
24:06
No. So the idea actually came about with, with ownership as we were looking for a way to
24:12
standardize and simplify our, you know, an OTB airport or a daily recap reporter is kind of
24:21
okay, I can see value, but now how much more can I get out of this? So what if I took this
24:27
information from the DMS and CRM and pulled it in to have one centralized report, but now if I
24:34
could do that and layer on marketing data, analytic data, results data, now that just makes
24:41
this even more powerful for us to be able to utilize and scale to a whole different level.
24:47
And then why did you decide to create your own tool versus go to someone else? And then I'd love
24:52
to have you walk us through what the tool is and then how can a dealer practically do this today?
24:56
But why, why do it yourself? Just for to your point where the concerns of data privacy and data
25:04
health data to me is one of the most important things. And I feel that is one of my most important
25:11
jobs is to protect that information, protect that data. So I look at that as my, you know,
25:16
the Brinks armored truck driver, that's me of the dealership and nobody's getting into it. So
25:23
I make sure that are you are you the AI expert at your store then at the organization? Are you the
25:29
person people go to if they want to understand, you know, it's myself and, you know, the beauty of
25:35
is I have Mike, our marketing director, he he's the creative side of it. So he uses it so much AI
25:42
for creative for script writing for, you know, overlays for everything we need on that side.
25:50
And we kind of work in tandem on a lot of different things, but data information analytics,
25:56
that's kind of my bread and butter. Okay, so I'm a car dealer. I'm watching this show.
26:02
I've heard chat GPT. I've heard you talk about the solution and the benefit.
26:06
What can I do to create something similar to what you have?
26:12
So first and foremost, sign up and you have to use the paid version of chat GPT. And from there,
26:18
you can get API access. Okay. Step one, step two is looking at your current text,
26:25
CRM, DMS, those are going to be the ones that take a little bit time, but they will have APIs that
26:31
are open and that you can use to query. Wait, so now, now just backup. So that means if I understand
26:38
you correctly, you do the paid version, I think it's like 200 bucks a month. I actually saw it
26:42
last night when you play, right? Yep. And let's say CDK, he leads whoever. If there's an integration
26:50
into CDK, I can actually link this paid protected version right into the DMS and it will call for
26:56
information. Okay. Close. Okay. So that's step three. So that's step three. You get excited,
27:03
which I love. I am excited because somebody is about to explain what the heck is going on.
27:08
Like I have not understood it yet. Everybody's like, Oh, it's so easy. Just do this. That I'm
27:12
like, wait, it's not that easy. Otherwise, everybody's doing it. So go ahead. Step two.
27:16
So step two is, yes, reach out to, you know, CDK, which they have their Fortellus
27:23
integration API cloud, which you can have access to DMS and CRM data.
27:29
Yeah, you get approval from them. And now you have that integration. Step three is you create
27:34
your e-store, right? Yep, your e-store. Great integration. Yep. Yep. And then step three is
27:38
you're setting up your AWS data clean room. So your data clean room for everything from
27:44
CDK, from everything from your VIN solutions or Techian or whatever information you're
27:50
feeding it, you set that to run to your AWS data clean room. Who do you call it AWS to create
27:58
up this data clean room? Is this the same thing as a CDP or a data lake? Or is it different?
28:02
Very similar. I do, I was looking at the angle as we're in CDP discovery conversations that we may
28:10
transfer over to that instance, but right now I needed a solution. So I just went on AWS and then
28:17
I use ChatGPT to, hey, how do I use and set up a data clean room on Amazon web server? And it
28:24
gave me 40 pages of information, instructions, how to do it. Wow. And then I connected the APIs
28:31
from ChatGPT to the clean room, to from CDK's API to the clean room, and then Google Analytics,
28:42
MetaBusiness Manager, et cetera. So all that information now is utilizing everything that's
28:49
protected in one instance and not pushing it out. So how are you comfortable that as you put it into
28:57
that clean room? Who has the liability if there's a data breach out of this quote unquote clean
29:03
room? Is it AWS? Is it you? Who is it? No, that'd be AWS and that's part of their security elements
29:09
of it. Okay. All right. So now the API is live. And in automotive, we all know APIs from all the
29:18
integrations to the CRM to the DMS and FNI, right? It's that crazy thing that fails a lot on a
29:25
Saturday, last Saturday of the month, and you can't call contracts or pricing or print anything
29:29
out, right? When you forget to, you know, decode and put the county in and you can't push deals,
29:34
yeah, absolutely. Yes, yes, yes. So everybody has familiarity with APIs. All right, what's the next
29:40
step? So then it's, it's trading. So then it's running different questions against it and starting
29:46
measuring results. And then it's enhancing that API through GPT of giving feedback. So when you
29:53
have that, and you pay for that API, you can provide feedback of this was wrong, this didn't,
29:59
you know, pull it in correctly, et cetera, you start feeding it back information every
30:03
single day to make it better and better as you grow. So that would be my other question, Jared,
30:09
is there are all sorts of different formats, you're getting data from, let's say a CRM and a DMS and
30:15
from Google Analytics and different lead providers, and you're just putting it all in this space.
30:20
Like, there are different formats. They're organized in different ways. They mean different
30:25
things. How do you get them all to play well with each other? Different name, different naming
30:29
conventions, even. Right, right. So, so is that part of the AI thing that's just pixie dust?
30:37
Okay. No, it's not pixie dust. You have to tell the AI, you know, first name here in this field
30:44
is the equal to customer there's mapping. Got it. So that's, that's part of the training that
30:50
you're talking about. Not only you have to ask it the question, you need to tell it where to
30:54
find the answer and how to find the answer. Right. Correct. Before you can ask it a question to
30:59
provide data results to you, you have to take the time to go tell it to map first name from field
31:07
A to customer first name from field Z16 equal each other. So that's part of that first discovery
31:18
phase to it and it takes time and yeah, but the benefit's worth it. So as you set up that integration
31:26
with the API and you push the data in, does it just automatically prompt you for that or do you
31:30
have to tell it, hey, I need you to prompt me for this? No, I sit there and tell it what to
31:36
connect in terms of what's how to think about it. Yeah.
31:40
Well, it's just how to consume it. It's no different than if you think about it,
31:44
how we are trained to go read a financial statement. And when you're training somebody
31:49
at the store, okay, here's your advertising credits. Here's your advertising cost. This is what
31:55
that means. That's in essence what you're saying, and you're doing to the GPT is with this data
32:01
field. This field equals this. This equals this to itself. Yeah. What's the most surprising
32:10
as you did this and you did this for the first time, like you're you're experiencing and discovering
32:16
something that's not been done by you or anybody around you. What was the thing that was most
32:20
surprising? Like was there a takeaway where you're like, holy cow, that's cool. Like I never knew
32:26
that I was surprised that that could be done. Though actually, if something I found I was just
32:31
querying it yesterday or not yesterday, but last week, and it was something I was just like, oh,
32:35
I wonder what's our close rate of leads that are older than 21 days? Oh, I want to know these things.
32:41
The fact I don't have this drives me nuts. Okay, good. Right. So I was like, oh, this, you know,
32:46
this store on the hoe, we're only at 2% close rate on these leads. Okay. And now it's like 21
32:53
days in order. Okay. Now identify which lead sources have the highest and lowest engagement
33:00
after 21 days. Yeah. Okay. And now it's the situation of why. So then you can take this back
33:07
and I can have a conversation with that ecom director say, Hey, here's the data. Let's go
33:14
inspect this a little bit further. Is this a people process or a provider? Which one of those
33:20
is the problem here? So that we can identify it and make adjustments to continue to grow and get
33:26
better. Very cool. So what did this cost you all in? So the the integration, the AWS space,
33:36
the subscription, I think 200 bucks a month or whatever, what what what is the cost?
33:41
I mean, it ended up costing me in terms just, you know, integration cost and some software
33:46
costs ended up costing me right about $1,000. And then of course, you know, there's, there's
33:53
the ongoing pieces that, you know, subscription. No, just up front was about 1000 and runs
33:58
three, that's a while 800 to 1200 a month, based upon how much data we're feeding it,
34:05
how many people for all the rooftops, right? Yeah, individually. Yeah. And a lot of those
34:12
are paid for calls of the API. So if you do it a lot more, it's going to go up. If you don't do
34:17
it as much, it goes down. And who charges that? Is that AWS or is it the source of the data through
34:23
the API? So everybody, everybody charges for data. Everybody's charging. Everybody's got their hand
34:29
out. So I mean, you're paying like a toll. Yeah, it's exactly like a toll in every direction.
34:36
Yeah. But, but the chat GPT side is just a monthly flat, right? So it's, it would be,
34:42
but they limit you, but they limit you to how many calls of the, got it. Okay. That makes sense.
34:53
That makes sense. So Lauren asks a great question. She says, does it do things like help your desk
35:01
managers write better deals or your inventory manager better position your lot? She's asking that.
35:07
So it's funny. I was so funny. She says that I was with David Long's morning clubhouse room and I was,
35:15
somebody mentioned like, gosh, what if you just took all your FNI lender docs and uploaded to this
35:23
and you could literally have your FNI guys figure out based upon, you know, DTI and score and rank,
35:32
which this could almost make it so that this would just tell them, which bank is the best option
35:38
based upon the customer's dynamic and be able to give you a profitability scale of which bank
35:45
you should send it to. And I'm just, I'm sitting here thinking, I'm like, taking notes. I'm like,
35:50
okay, let's add this to future planning. So that's something I have penciled that I want to add to
35:57
it. But I think that'd be critical for the FNI team. It's one of those things where you don't know
36:05
what you don't know, right? So as you're going through this seemingly limitless configurability
36:10
of this system, it's like, you can really get down to the nitty gritty. It's, it's really
36:16
remarkable. I mean, I'm curious to see what questions you come up with in the future and how
36:21
you can, you know, action that out. Yeah, no, myself as well. So, so it's interesting. Why do you
36:30
think so many of the AI providers charge so much for their technology right now? So I always give
36:37
the example of one of our, a store out there that paid a $4,000 setup fee, $4,000 a month,
36:42
it was a bolt-in to their CRM, they had no idea what it was doing, but they paid it just because
36:48
it seemed like cool tech. You set this up for $1,000 and you've got some nominal, you describe it,
36:55
monthly fees. Like, why are some of the providers paying so much when you can do so much on your own?
37:01
I think it comes down to, you know, time, but it also comes down to our mindset and automotive.
37:09
We are very reactionary and we are also very focused on the 30-day window.
37:15
Yeah. So it's one of the things is I need that shot in the arm now. This something can help me
37:20
sell 10 more cars a month. Yeah, I go pay three grand for it, but I'm selling 10, 12 more cars or
37:26
I'm solving for something like, you know, an overnight lead volume and hey, I'm getting 60%
37:34
of my leads overnight. I have to have AI to be able to respond to these after hours or else I'm
37:40
going to get hosed because the two stores that I'm in competition with have something, I'm never
37:46
going to get that engagement. So it's, I think that's the difference too, it is this is help
37:52
operational AI. Yeah. This is more operational kind of leadership management type of things
37:58
when you're mentioning it or tools that are generative and conversational AI tools that help fill
38:06
a void and or gap of our day-to-day operation. Yeah. Tyler LeBlanc says, when you talk about
38:13
that FNI tool, SelectFI does that, right? So SelectFI apparently does the feature that you're talking
38:20
about relative to providing finance feedback. So if you were to, and we've only got a couple
38:26
minutes left, the producers are saying, we got to go, we got to go, we got to go. But I just want
38:30
a couple more questions, such an exciting topic to me. What is, what would you do differently if
38:36
you were building this again today from scratch? What's a takeaway that another dealer wanting
38:42
to replicate what you've done could learn from? I wouldn't have waited. I would not have waited.
38:49
I mean, I waited probably six months after just initial thoughts to do it. I would have started
38:57
right down right there, even if it's simply as starting with the CRM or DMS integration request,
39:04
because that does take a little bit of time. Yeah. So, and that could slow you down and then we get
39:10
distracted and now it's, you know, by the wayside. So biggest thing I would say takeaway is if you're
39:17
thinking about it, start now, make a plan, draw a plan out of where you want this to be in a year
39:24
and try and get from one year to six months and then continue to grow out.
39:29
I love that advice. So it's with GMs right now. Who's your next rollout? This will be our last
39:35
question because they're driving me nuts to go. No, I love it. So the next piece I'm going to be
39:40
rolling this to is our ECOM team for a better accountability across the day-to-day operations.
39:47
Yeah. Yeah. Very cool. Well, I got to tell you, this is a fascinating conversation, Jared Killaway.
39:54
You've explained things to us that everybody wants to know with specific steps on how to
39:59
achieve those next steps, which is awesome. So VP of Digital Ops at CAS Auto Group, Jared Killaway,
40:05
thanks for joining the show and thanks for sharing your perspectives on all things chat,
40:09
GPT. By the way, a lot of the prompts that were in the CDG email and newsletter that went out came
40:16
from you. So those are tried and tested in that environment as well. So Jared Killaway, thanks
40:21
for being here. Thanks, Sean. Oh, it's a pleasure. Thanks for having me so much.
40:27
And guys, now a quick note from our sponsors, Toma. So most voice AI is frustrated customers
40:34
and waste everybody's time, but only Toma customizes AI agents to your dealership.
40:39
So calls actually get resolved. That's why top dealer groups trust Toma to automate their phones.
40:45
Visit www.toma.com forward slash CDG to learn more. And I want to do the same thing here.
40:52
If you can figure out how to use your phone and scan the QR code here, you'll be good to go. You
40:57
can hop right in. So we just want to say thank you again to Toma for sponsoring the show and
41:02
letting us do, letting us do what we do best. Giving, giving voice to all of automotive. That's
41:08
awesome. That's right. So all right. Next, next up, let's dive into this today's industry pulse.
41:16
Of course, there has to be a cool graphic for that. There it is.
41:25
All right. At Sawyer Merit on X posted something that got a lot of traction over the weekend.
41:32
He claims that all the cars on the market today, what do the best, what's the best selling car,
41:37
the fastest production car, the safest car and the most American made car have in common?
41:42
So think about this. What do all these cars have in common? Best selling, fastest production,
41:48
safest and most American made car have in common? I don't know if I'm going to agree on this, but
41:54
we'll go down this rabbit hole. They are all EVs, which got us to thinking then why are they so
42:01
hard to sell? That's so true. I don't think they're hard to sell. I just think politics has pushed
42:09
demand. Politics has tried to push EV demand beyond what it naturally is and infrastructure
42:16
hasn't been there. Consumer understanding, education and appetite hasn't quite yet been there.
42:24
Then battery technology in the past hasn't been there. I'm of the belief, Julie, that as the
42:32
incentives peel away, we're going to get to some sort of a natural demand. It's going to be different
42:37
all across the country based on geography, how much infrastructure there is in different
42:42
geographies, but there will be demand and it'll probably surprise some people where that sits.
42:47
The big mistake and I think, I mean, this is politics, but the big mistake was
42:52
government tried to create demand where there just wasn't demand and they tried to force
42:57
infrastructure where there wasn't infrastructure and it was costly. And by the way,
43:01
nobody wants to have this debate, but they've made that mistake during things like COVID and
43:07
create scarcity, don't create oversupply. When you want to push out an initiative,
43:14
create exclusivity and create scarcity. Don't flood a market. When something can become too easy to get,
43:24
nobody's willing to pay anything. It takes me back to my English roots as an English major.
43:29
I forget what the quote is, but whatever we can get too easily, we don't value. And that's the
43:34
case with EV. That's the case with a lot of different things that government does. Government
43:38
is a very blunt instrument in the marketplace. And listen, and it's also not a one size fits all
43:45
approach. I mean, this was a very easy or rather relatively easy adoption for Europeans because
43:51
the way society is set up is so completely different. We live in the United States of America.
43:56
We have a vast amount of terrain to travel every single day. So I mean, EVs just don't fit
44:03
everywhere. Which now I think why it's so exciting to see this evolution, right? The plug-in hybrids
44:09
and now the, what are they called the ER EVs? The EVs with gas generators. Like, that's it, man.
44:16
If I can drive around and bring a charging station with me, that solves the whole EV
44:21
adoption problem right there. All right. Well, that's a wrap on today's industry
44:26
pulse. Fascinating conversation. Thanks for the post to X. You can join the conversation after
44:31
today's show and post comments as you want, just as a segue into our next segment. Actually,
44:37
we'll cover these comments. A lot of great comments online coming from our audience.
44:42
We'll recap some of them at the very end. But let's head to our next guest,
44:45
founder of Auto Genius, Ben Haddon. Welcome to the show. What's good? Welcome, Ben. What's happening?
44:54
Ben, thanks for being here. Doing great. This is so awesome. Great. This is awesome. Thanks for
45:00
being here and having a voice here in the voice of automotive on this Daily Dealer Live show. So
45:06
before we get into it, I'd love to ask you two questions. How's Biz? And as part of that, tell
45:10
us a little bit about yourself and what you do at Auto Genius. Oh, business is crazy. Let's see,
45:17
our company's grown about 400, 500% over the last nine months. We have a little bit of an
45:25
interesting story. I've been in auto for since like 2008 about early dealer.com had a bunch of
45:35
startups in between, sold them. And we made it a mission to make a startup that was guided by a
45:42
community. So we created this thing called the Auto Genius Community, like 6,000 dealerships now,
45:50
or represents about 6,000 dealerships now. And they kind of, they're our compass for our product
45:55
roadmap. And the major problem that we found that nobody could solve was speeding up websites. So
46:02
we figured out how to increase the speed of any website on the internet by 50%.
46:07
And that increases your leads 34%. But not really what I wanted to share today.
46:14
What I have for you today actually has a lot to do with what you guys are just talking about.
46:18
All right, how does it relate? So I have a Rivian, I had a Model S, right? My family actually,
46:27
see my uncle has four dealerships here in Burlington, Vermont.
46:34
And what I've continually realized through sort of like all the tech adoption cycles
46:42
was that there was sort of this rebuttal that kept on coming up.
46:47
Right when I was starting to like introduce websites into the market through dealer.com,
46:54
people would say, well, why would I want your website? Maybe just a webpage with my address on
46:58
it. Right? Because if they want to see something, they should just come on in. And then it became
47:04
like price transparency. Well, if they want to see that, they should just come on in. And then
47:08
it became digital retailing. Well, if they want to see that, they should just come on in. And it
47:13
was like, really, like, is this going to be like the rebuttal every freaking time?
47:17
Right. And then I realized, by the way, Ben, I get that before you say I realize, I get that
47:24
because in automotive, we've been taught since the early 90s when I got into the car business,
47:28
we're best when that customer comes in and sits in front of us. And so I get why old dogs push
47:35
back on that idea of turning it over to a more virtual space. So then your thought was, thank
47:40
you, Ben. I figured out and sort of started graphing together that a lot of dealerships
47:52
want to believe they're an optional purchase. But then they start behaving like they're an
47:57
obligated purchase. You can't have both. And an example of that, you know, if I was the only
48:04
guy that had water, I can just play a game where I say, I'll wait till they get thirsty.
48:09
Sure. A lot of dealers can play that game. Actually, my uncle here, I think his Audi store,
48:15
yeah, you have to drive like three and a half hours to go get an Audi in Boston,
48:19
like otherwise, you know, I got the water, man. You know, you want an Audi? Come on in.
48:25
But that also explains why dealers are really hesitant to innovate. Because if you got the water,
48:32
and you can just literally wait until people get thirsty, then why do I need all this tech?
48:39
Or why do I need you all these things? The aha moment for me, if you remember, I don't know,
48:46
if it was like 2016, 17, somewhere in there. Coxado had a study, it was like
48:55
customers hate spending three and a half hours in a dealership.
48:59
Yes, which we've known for a long time. Yes.
49:02
But here's how I would test your brain to see if you're a little too data drunk. Okay.
49:09
What did your brain immediately go and try to optimize in that sentence?
49:14
The time in the store. The number. Yeah. Right. We actually have gotten so short-term thinking
49:21
in automotive that what we immediately beam into is, well, I'll take that three and a half hours,
49:27
I'll make it 45 minutes. Our minds skip over the word hate, which is crazy, because we might
49:34
just make a 45 minute process people hate. In other words, there's no such thing as a fast
49:39
cruise ship, right? A cruise ship is slow as hell, and that is actually part of the experience.
49:46
If I went to go get a Hermes purse, the 18 months that I have to wait to get my wife that purse
49:53
is actually a benefit. She gets the brag that she's getting such a rare. Oh, I have to. Same
50:00
with Ferrari buyers, same with Porsche buyers. If you have a black Submariner on your wrist right
50:05
now, you could probably just walk into a Rolex place and grab it. The guys that brag about their
50:09
Rolex are like, no, no, no. I've got the ceramic Daytona. It took like 18 years for them to finally
50:17
get, you know, so then this strategy that you're talking about, this is what you call the blue
50:23
ocean strategy. And this is an idea of creating a space that is uncontested. It's taking it above
50:30
price. It's taking it above all the traditional constraints of automotive, and it's creating a
50:35
space where the value delivered to a buyer or a consumer is so high that they become comfortable
50:43
in that space, even in the face of different competitive facets, different marketplace facets
50:50
and all that. So when you think about creating that space, what are some of the steps to getting
50:55
there? Like to a dealer listening today saying, you know what, I get that. I see the need for
51:01
a space where I don't have to compete on price, where I don't have to compete on availability of
51:06
a particular vehicle or features. I want to compete on value and delivering an excellent
51:12
customer experience. So the first thing I think to figure out is where are you
51:19
on the axis of optional to obligated, right? So, you know, my uncle actually really can just
51:30
play in this obligated space, right? Like he can absolutely hang out there and crush the market
51:37
because the market has basically been dominated and I'm not going to use the word monopoly,
51:42
but he's got your own market, right? Now, if you're in a metro area with like 20 competitors around
51:49
you, then as soon as we switch from a customer saying, well, I'm obligated to buy a car because
52:00
I live in America. And how that kind of has to do with the EV thing is that a lot of people
52:08
hate their obligations and an EV is obligating me to also get a charger and it's obligating
52:14
me to stop at a place for 25 minutes to charge it and it's obligating me that it's actually adding
52:19
net more obligations. It did a raw effect. 100%. The brilliant thing with Tesla to the point earlier
52:26
was they made it the wait list, right? They actually, that wait list became a benefit.
52:31
They created scarcity. That's my whole point. Yes. 100%. You create scarcity. If the government
52:37
had created scarcity around EVs, people would be clamoring for them. There's certain health,
52:42
there's health initiatives that if there'd been scarcity around being able to get this thing or
52:47
that thing, people would have paid a lot of money to go get it. And that's this space you're talking
52:53
about where you don't have to compete on price. You can compete on the value delivered by it.
52:59
And by the way, neither is right. Like what I'm saying is I'm suggesting there's multiple ways
53:06
to exist as a dealership. We've all known that. And what I've kind of created is a map to get you
53:12
into the strategy or reflect the strategy that you want to find the blue ocean. Your blue ocean
53:18
might be experiential. Cool. We can get you there. Your blue ocean might be more towards
53:25
instant gratification. Get you there as well. So if I'm a dealer sitting watching this and I'm like,
53:32
okay, I kind of get what you're talking about. Some of it's like nebulous, right? What are some
53:38
specific steps I could look for to figure out that blue ocean for my own business? And how do I
53:44
pull up the graph space? Yes. Producer Sam, could you please give us the graph? And by the way,
53:50
to our audience, don't freak out when you see this graph because this graph is very busy. But go
53:54
ahead. It's crazy. But what you're looking at is the magic quadrant. And you're looking at the
53:59
bottom portion is optional purchases. Okay, as you go up to the top, you're looking at obligated
54:05
purchases. The far left, you're looking at something that's delayed gratification. And the far right,
54:12
you're looking at something that's instant gratification. If you can zoom in on the top left,
54:19
what you'll see is car buying. Yeah, I see it there. Car buying is surrounded by other things
54:27
that are slow obligations like replacing your roof, going to the dentist. All of these areas
54:33
actually have insanely predictable cash flow. PE companies seem to love them for roll up strategies
54:40
because they can then innovate them pretty easily
54:45
and compound the growth. If you look at where Carvana is, which is not too far away from the
54:53
regular car buying, it's just drifted a little bit towards instant gratification. That's them
54:58
seeking out blue ocean. But if you look down towards where Ferrari or Porsche is,
55:06
which is down in the optional, like you don't need a Porsche, right? That's starting to bring you
55:13
into more experiential. And so as you go either direction, what you'll also see are marketing
55:19
channels that help you get to that direction. So if I'm trying to be more like Porsche,
55:25
then I'm having events at my dealership. Events are a very optional, slow thing, right? If I'm
55:35
trying to be like more on the instant gratification side, then I'm going to be doing TikTok. I'm
55:43
going to be doing Instagram Reels. I'm trying to get people that are impulsivity-based buyers.
55:49
So each one of these, actually, you can draw a line from where you are to the marketing channel
55:54
as a path to get more and more of the blue ocean that that area would capture.
56:00
So everybody sort of starts in that top left area, but then can find little pockets where
56:06
they can go, you know what? Like I know for a fact right now, the undervalued mediums
56:14
are probably in the bottom left in traditional, which is crazy because I've been a digital guy
56:21
my entire life. But what I realized is like no one would send a wedding invitation through email.
56:31
Right. Right. Like if you have something important to say, you would use direct mail.
56:35
You'd use the slowest, most optional thing to get in front of that person, right?
56:45
Is that right? I don't know. Debate it. I love it. I'm not sure. I'm not sure that is correct.
56:49
Like any more, nobody goes to the mailbox to pick up a piece of mail. So I think wedding invite is
56:55
a very specific one, right? Aside from that, there's a lot of junk that comes in the mail.
57:00
I never 100% of it is a lot of it is still meaningless. Yes. Yeah. Yeah. But if you actually
57:07
showcase effort in in your direct mail, it'll it'll work. 100%. I'll ask you guys a question.
57:15
Like think about this. You have a million dollars to spend right now. You go spend $500,000 on
57:23
producing and like getting a great cast together, making a beautiful ad.
57:27
Then you spend the other $500,000 and you can either spend it on TV or you spend it on OTT.
57:34
Yep. Right. Or YouTube pre-roll. If you're a viewer, which one inherently feels like it has
57:42
more meaning, even if it's the same exact ad? Julie, what do you got? Yeah. OTT.
57:50
Hell no. What are you talking about? Wrong. If I saw a Doritos ad on the Super Bowl versus a Doritos,
58:00
the same exact Doritos ad on YouTube pre-roll, digital pre-roll. Yeah. Yeah. Exactly. Digital
58:08
marketing inherently actually feels cheap, right? Digital marketing is almost the way of saying,
58:13
I subconsciously am telling the consumer I paid the least amount of money to get in front of your
58:18
eyeballs. But when it comes to effective connection with that consumer, the way digital media can
58:24
saturate and give a message over and over again, to your point, it may feel less, but it's actually
58:29
more effective at delivering the message. Yes, no. That's got to be our last one because we got five
58:33
seconds less left in the show. What would you say? Shoot. I would say no. I would say saturation.
58:42
So now you're looking at saturation versus quality. Yeah. I would say depending on the
58:47
strategy of the dealership, if you're trying to just hit as many eyeballs as possible, fair.
58:54
But if you're trying to generate meaning, which is going to be more important in an AI world,
59:00
like an AI world starts delivering you things instantly, but at the same time, you are,
59:07
yeah, exactly. I want to be different from everybody else. Ben Hadley, I hate to cut you.
59:12
Founder of Auto Genius. It's been an incredibly interesting conversation. Thanks for showing
59:16
us how to seek out those areas where we can deliver the best value to our customers
59:19
every single time. Appreciate you being on the show, sharing your perspective. Yeah, thanks for
59:22
having me. Thanks, Ben. Dude, we had a lot of great conversations just going in the last 20
59:28
seconds here. YogaCar said most cases car purchases is an emotional purchase unless the car totally
59:33
broke apart. Lauren Klein, thanks. We are all old car guys. And then of course, Lauren Klein also
59:40
says the operational possibilities are endless if you're using that data correctly. And I think
59:45
that is, that is true, right? Yeah. But good show today. And we learned something about ChatGPT.
59:52
I'm going to go out and research AWS, like the sterile room for protective data, because intuitively
59:57
I've known I don't want to put data out there for sure. So that is, that is smart. And to our
00:02
Daily Dealer Live audience, thanks for watching Daily Dealer Live. We break down the biggest moves
00:06
in the car business as they happen. Don't forget we're here live every Monday, Wednesday, Friday,
00:11
1pm Eastern. So if this is your world, hit subscribe, like, turn on those notifications
00:16
so you never ever miss a beat. And we'll see you next episode Wednesday, everybody. We'll see you