Listening To 4000 AI Phone Calls with Monik Pamecha

Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Paul J Daly (00:00):
And you guys know good riddle. No,

Unknown (00:09):
this is Auto Collabs.
Remember

the riddles you try to solve in like middle school?

Kyle Mountsier (00:14):
My dad, my dad literally. And this is one time
he came to our house and we werejust sitting on the couches,
hanging out, talking, and hejust started telling riddles to
my kids, to, like, discover andit was an endless Trevor
treasure trove of riddles overand over and over. And I was
like, where do you store

right, right, where's all this coming from? It
was wild.

Michael Cirillo (00:37):
I grew I feel like I grew up in a riddle. You
know what I mean? It's like,like, like, my dad, I get it
now. I get the whole thing.
Yeah, it's like, dad walks inand says, You have four sisters,
and that's the that's it. It'slike, figure that

out. Well, then, then you're like, what? I only
know three of them. Wait aminute. All right, here I have a
riddle for you. I'm built inlines, but sold in rows. I carry
people where they go. My heartis fire. My breath is there. I
sleep in lots, but roameverywhere with four feet
turning. I dash with my What amI shining in the light? This

is Michael Cirillo at 3am with ferocious
heartburn.

It's a car to car.
Yeah, I don't know that waspretty lame. GPT,

Unknown (01:27):
US GPT for that. Well, well, yeah, it's built

in assembly lines. And, yeah, it

was real easy. It's it wasn't that. It definitely

Unknown (01:39):
wasn't hard cars.
I know. I know. Well, todaywe're get to, get to meet a new
person. Monik from Toma, talkabout solving riddles. What do
you do with so much information,especially with phone calls and
all the data being produced? Howdo you make something useful out
of it? I mean, more useful thanthe riddle I just read. We hope
you enjoyed this conversation wehave. Was you gonna say

(02:00):
something. Who two saidsomething? No, I thought I had
chair. Was that? What it was,we'll blame it on Kyle's chair,
either way, we hope you enjoyedthis interview with Monik.
Monik, thank you so much forjoining us today. It's so good
to be here with you. Hey, Paul,it's my pleasure. Alright, so

(02:22):
one of the favorite things thatwe love about this particular
show is that we don't just talktech and talk strategy. We like
to talk about the people side ofthe business. And one question
that I personally love askingis, how did you actually get
into the auto industry?

Monik Pamecha (02:38):
Well, surrender pity. I feel like anybody else
I've spoken to who's been in theauto business always told me
that, hey, the car businesschooses. You don't, you don't
get to choose. And I think it'sbeen true. I just somehow landed
in there, but it's been a funjourney. You know, my background

(03:01):
is like, actually engineering,building things, selling things
like that had nothing to do withcars for the most part, just
software. And then we, we just,we were working on a project
with my co founder for a while,and we did a lot of different
things, lot of random things,like we built stuff for diet
recommendation tools forpatients with chronic

(03:23):
conditions. Uh, my co founderactually sold nipple covers for
a while on Amazon. Uh,lot of learnings, but eventually
we started building somethingthat we had a we had one dealer
that really liked what we did,and so we were building a voice
AI platform for them. And theythen went and, you know, spoke

(03:48):
to their 20 group, and they toldthem that, hey, this thing has
been kind of working. It washorrible at the first when we
made the first version,

you're like, don't tell anyone, please.

Unknown (04:02):
Well, they say anything worth doing is worth doing
terribly at first. Yeah,

right. And it was so funny, like, if I start
talking about the bugs and theissues we had, you laugh. Like
we made an error, wheresomewhere we left a comment
saying, act like a cowboy, andwe forgot to remove it. And the
next thing is, the call goes in,and it's like, how do you
partner? Like, would you like tobook an oil change? And the
person actually goes and booksthe oil change, but I have the

(04:33):
actual call recording. It'sreally funny. And the general
manager then email me and say,hey, the AI is acting like a
cowboy. And I'm like, Yeah,because that's what it's

Unknown (04:42):
a cowboy who can close is the only thing you need to
worry about, right, apparently.

But it was all funny, like that, right? Like,
sometimes it take, like, it waslike, talking to a walkie
talkie, you know, you'd get aresponse in like, 10 seconds,
and people are patient,whatever, right? Like, we just
started with that. I. We had onedealer that kind of that that,
you know, was okay with all theproblems that we had, still
liked it. And told the 20 group,and the more folks came into us,

(05:09):
and they're like, looks like,you know, what you have is, you
know, something that solves ourproblem. So maybe you should,
you know, help us too. And thenwe said, why not? Let's try it.
At that point, we had no ideaabout, like, why dealerships
want this, like, we because wewere working with hospitals,
with, like, banks, and everybodywas moving so slowly. But

(05:31):
dealers were like,

hey, what go? Let's go. Wait a second. We'll test
this sucker tomorrow, right?
Let's say

that again.
Because I think, Well, I think,you know, dealers are a lot of
times seen from otherindustries, and like retail Auto
is seen as, like, the slowstream behind. But you're like,
No, no, well, and it sounded youwere working with a lot of
regulated industries, which arealready somewhat slow. But of
those, at least, it seemed likethe dealers were kind of like,

(05:59):
ready to throw in a little bit,

yeah, like, because a bank gave me a
questionnaire, and they said,how can you guarantee that this
AI is not going to, like, youknow, be informal. And I was
like, how can you guarantee ifyour agents are not being
informal at the call center,right? Like, that's a really

Unknown (06:24):
tough question. That's a great, that's a great. Like,
how do you guarantee it's like,well, we monitor them. It's
like, Yep, so do we?

And, I mean, like, how much formality do you
need in, like, Home Loans? Imean, I'm sure there's like, you
know, obviously the regulations,different countries have
different rules, yeah, Howdy.

Howdy partner might not do it,

Unknown (06:42):
right? Yeah, right.
Give

me my social security number.

But obviously, in its infancy, like, you know, the
technology is new, there are alot of risks associated with it
has matured a lot over the lastyear, so these questions don't
even come up anymore. But, youknow, a year and a half ago,
that was like, that was how itwas, and dealers were willing to
take risks. And personally, Ilove that so much. Because if
you're building something newand you really wanted to break

(07:07):
new ground, and like you know,you want to explore new
territory, you will fail, ofcourse, right? You'll make
mistakes, but you need to havethat space for failure. We don't
have any of that with otherindustries because they're, you
know, little slow moving, right?
For a lot of reasons. That's whywe loved it. And when we step
into the dealership, like, wewalk around and we talk, we
spoke to like, every person,right, from the person who's the

(07:28):
lot Porter to like the generalmanager, owners, every person in
the chain, and we're like, like,everybody had, like, all of
these ideas on like, you knowhow this should be different,
and sometimes it's conflictingwith, you know, what other
people are saying in the storetoo, right? Which is pretty
natural. That's when we startedseeing, like, you know, common

(07:48):
thread, like, okay, there is alot of opportunity, and it's
very underserved, despite therebeing so many, like, tools, I
don't think they're reallyhitting the mark. And, like,
solving all the pain points,some of them do a really good
job, but like, overall, there'sstill a lot of scope. So all in
all, we're like, We're buildingsomething. There's some
traction. There's so much more,you know, that we can do, put it

(08:11):
all together. We're like, Okay,we're going all in like, forget
the banks, forget the hospitals.
Like, I don't know how manydecades it's going to take to
make it something reallyexceptional over there, but here
we're just going to dive in anddo our job,

and you're never going back, never go. It's not
possible. Just so, you know, youcan't get out. You're locked in.

Unknown (08:37):
It chooses you, and then you're in. Yeah,

so we were talking that you're a soccer
fan, yeah, before this, here wego. And so I coach, I know, I
know, Paul, sorry. Just hangwith us for a minute. Paul, hang
with us. So unfortunately,Monik, he's also a Real Madrid
fan, which is a problem, butwe'll get over it for a second.
Okay, I'm a city fan. All right,I have no problem. With either.

(09:05):
Alright. So, so I'm a soccercoach. I coach kids, and I've
been watching a lot of likesoccer coaching videos. I'm
going somewhere with this, but Iwas watching the soccer coaching
video, and this coach wascoaching probably like 12 year
old boys and girls, and theykept saying, Sorry, right? Like,
it's actually, if you knowsoccer, if you've been around

(09:26):
soccer, there's this very like,it's one of those words, right?
Like, unlucky, that people willsay. People will be like, Oh,
unlucky, you know, Oh, so sorry,miss. Like, if you missed the
pass or something like that. Andhe said, he said, Never, again.
Are you allowed to say sorry?
Because all that, all thathappened when you missed the
pass, when you were when youwere unsuccessful, is an

(09:48):
unsuccessful attempt atgreatness. Yeah, right. And so
like, keep attempting that.
Because, like, if you, if youwere attempting that. Pass into
the open field, and the personcould have run onto it, but you
just barely missed it. You don'thave to say sorry for messing up
because you didn't do anythingwrong. Sorry is wrong, right?

(10:11):
Unsuccessful attempts atgreatness are just unsuccessful.
And I think what you were sayingthere, and I think this is so
key, especially when we thinkabout like dealers trying new
things, or tech partners tryingnew things. A lot of times, we
just want solutions that aresilver bullets, that always work
100% of the time, and they neverfail, and we always know that

(10:32):
they provide value. Butsometimes there's unsuccessful
attempts at greatness that ifyou stop those, if you don't
give yourself the opportunity tomake those, then you never make
the cool thing next is theresomewhere that, like point back
to something where you're like,we've really tried this, and it
was an unsuccessful attempt atgreatness, but we know that,

(10:53):
like on the other side, there'sprobably something really cool.

So just to, I guess, to to back up that point
that you said, right about,like, trying things and things
not working out, and same thingwith, like soccer, right? Like,
I also that people saying sorry,I think that's like a more
Europe. Yeah, it's like Europesaying sorry, that's not

or Canadian. Yeah, little both.

But I think, like, when was the last time you
know something in your life andaccording to plan? Like,
everything just went accordingto plan. When was that one time
that happened? You know, I don'tremember any single time in my
life, like, whether it's inbusiness or outside of it, like,
so, I mean, that doesn't meanyou don't make plans, right? Of
course, you'll have plans andyou have some goals. So we

(11:44):
always, like, with our product,or, like, you know, many things
in the past, as well as I'vebuilt, like, you know, small pro
we, I used to build a lot oflike social networking websites,
like, in 2007 that was when Iwas in high school. And we're
trying to build all these likeproducts for users. And we would
build these amazing features.
We'd add like, Oh, if you're inCentral Park, you can make these
groups based on sports, andpeople will come in and you'll

(12:06):
like, all know when you're gonnaplay, and all that sounds so
great, right? It's what was thatapp called? It was called active
life NYC. So this was, I was inhigh school. I think I was like,
maybe 16 years old or 15 yearsold, and I found a customer. He
was a instructor in CentralPark, like, I think, a fitness

(12:28):
instructor, and he was like,let's make a social network.
Let's get everybody to, like,come here, and then, you know,
have these common interests. Andthen they'll put videos on their
photos on here, it'll be like amini Facebook of like, people
who just play sports, right? Andon paper, it's like, it's great.
You know, of course, you want tohave all these communities. You
want to know who you play with.

(12:49):
You want to come and do allthese things. We put all the
features in there. Every singlething. Like, I spent, like, a
year building it, and I thinkonly four people used it. Wow.
And it was, like, rough, becauseI spent so much time building
it, and I thought every singlefeature, I was like, this would

(13:10):
be amazing. This would be blah,blah, blah, and you're trying to
do all that, but, you know,people just didn't want it. They
just rejected it. And they'relike, Screw it. I'll just post
the picture on Facebook, forexample, right? So I think
throughout, like, even with ourcurrent product, we've tried,
like, if we have five featuresthat work for that, we have
like, 20 features that did work.
So every single thing, like, Iwould say, we've, like, done

(13:34):
that again and again. We try tothink ahead and get there. But I
think failure is just a naturalpart. And if you're even tired,
to say, sorry, we're like, youknow, whatever, it always
happens, right? It's accepted.
So Screw it. Let's do the nextthing.

So there's so much, there's so much, I think,
there's such a variety ofopinions and thoughts on what AI
actually is and how it canactually be used inside the
dealership to to cut down onrepetitive tasks, or to
increase, increase, leadquality, close rates, all of
these things, you have a uniqueposition in that you get to look

(14:08):
inside all of these things,formulate a product around them.
What do you see, as we'll callit, the lowest hanging fruit of
auto dealers who are thinkinglike, how do I implement AI into
my operations in some way. Whatdo you think is the the lowest
hanging fruit that you wouldencourage dealers to say, hey,
dip your toe in the water here.

I think if there is any specific thing that you
do that does not requirecreativity, it should be
automated. So, for example,right? Like booking appointments
for like, specific things, likeincomes, like whatever you know,
my whatever I need to requestfor, and then out goes specific
static appointment, automated.
But if there's like, somethingthat's more requires more

(14:52):
creativity, right, where you'reconstructing, like, some kind of
an offer you're trying to, like,appeal to someone that's better?
How? Handled by humans. So Ithink it's like first step is
distinction between what shouldbe automated and what should be
not. And then whatever has to beautomated, you automate that
part. So I'll tell you, becauseI've been listening, I've
listened to over 4000 calls,like maybe on 3x 4x feet over

(15:15):
the last year, year and a half,and a majority of the calls are
actually repetitive, like theyhave these elements of, you
know, like, even withinappointment bookings, when
you're dealing with specificloan or cars, like, there's
always, like, this rule in mind,you know, like, Oh, if this,
then that, if this, then don'tdo that, right? So a lot of it

(15:35):
would be even things that wedon't think are, you know,
repetitive. They actually are,because when you look at like,
you know, four or 5000 calls,

realize how much it actually happens

same thing you're just doing, like, if and else,
right? So a lot of that can beautomated. I think best use of
AI is actually doing whatevercan be done multiple times.
Like, you know, giving it to theAI and only taking the stuff
like that's really, reallyrequires you to involve yourself
and have that creativity. Andalso, like, you know, the
customer expects that, you know,because, I mean, what personal

(16:09):
touch Do you want when you'rebooking an oil change, right? I
mean, of course there is, tosome degree you want that. But
like, wouldn't you better servethat time? Like, you know,
giving, you know, pre qualifyingsomebody for a used car, you
know, things like that, right?
So it's understanding what toautomate, what not. And once
you've decided on what toautomate, going all out on that
end to end, and then taking therest and then giving it to

(16:31):
someone who's like, you know,trained for that purpose, right?
A human, yeah. Do

you get Do you tinker with automation outside
of, like, the work thing. Or doyou get tired of it in, in
building a product?

He's all real life is full manual, right? Like,
even as manual windows?

Yeah, exactly. I mean, let's see. I think the
Waymo so as some degree ofautomation outside, you know,
where it's it's it's actuallyreally funny, but, yeah, waymos
are cool. Like, at least in SanFrancisco, you can take them now
and then, initially it's like, alittle scary, but then once you
get in it and you're like, okay,like, if it says three minutes,

(17:12):
it'll be there in three minutes.
And you know, one thing is thatif I'm late, I'm always very
nervous, because, you know, ifI'm like, minute or two minutes
late, the Uber drivers. Like,just getting annoyed, but with
way more, I'm like, whatever,you know. And also, if you're
driving, you can always cut waymore off. Like, I mean, you
know, it's that aspect too, ifyou're on the other side. I

never thought about that, but oh, that's hilarious.
People just like, oh, yeah, it'sa robot.

Just like, Oh, it's fine. Yeah, readjust, yeah.
Way more on the other side,Elliot turn left, you know,
like, whatever.

Oh, that's hilarious. That's amazing. But,

yeah, not, not too much. I mean, um, just with
work, like, you have all theseemail automations and, like, you
know, summaries of meetings thatyou can get after that. So you
could do a lot of, like, smallthings that I've been doing,
they're pretty good. It'll justmake a to do list after every
meeting. So you never have to,like, think of like, oh, I need
to do this or that. But, uh,yeah. Like, randomizing what to

(18:14):
eat. I mean, some people love,you know, choosing what to eat.
For me, it's like, I'm just, Ikeep getting confused, you know?
So I have this extension, I'lljust randomize this thing on
DoorDash, and we'll just pick,like, all right, no way, item,
and then just get it, you know,

that's amazing.
You're like, I'm sick of makingdecisions. I'm sick of that,

and I'll eat it, you know. I'm like, whatever,
right? It is, what it is

that's but, but that's a great like, that's,
that's actually similar, butit's the same type of thing as,
like, Steve Jobs only wearsblack shirts, right? But to
reduce the decision makingnecessary so that he can make
the decisions on the hardthings, right? And it's, it's
the same thing. It's likedeciding what food today put

(19:00):
that on the robot like it'lljust pick something different
for me, I'll do the hard thing,which is listening to 4000 calls
to see, oh my goodness, youknow, to see if we've made the
if then statements correctacross that many calls. That's
the hard thing. That's whatneeds the human thing. What are
you most excited about in yourproduct as you're like, building

(19:21):
toward the next things, I thinka lot of people are starting to
get familiar with, oh, we canhave some level of AI or machine
learning, kind of, like, takeand capture, especially service
calls, because they're highvolume, they're repetitive. But
what are you excited about asthe evolution of, like, voice AI
when it comes to consumerinteraction.

So I think it's really voice, AI is like, just a
channel. It could be text, itcould be email, it could be any
channel, right? It's really thework that gets done behind the
receiver. Like, that'scomplicated. And what has
happened over time is that, youknow, initially, maybe 20 years
ago, you had a. Operator. Thenyou had a phone tree, right?
Which is like, press one, presstwo, press three, for this,

(20:04):
right? This is the nextgeneration, like, next evolution
of that, which is like, hey,what do you need help with?
Right? And then you can, like,figure out and do things. You
know, automatically, a lot oftimes you'll transfer in cases
where you cannot help them. Whatis happening is that your the
things that you have totransport for, will keep
shrinking over time. So, youknow, let's say the boss you'd

(20:27):
be able to handle only 2% of thecalls, then it became 10, then
it became 15, then it became 40,right? So I think we'll see that
number go, you know, like, like,how much you can handle will
keep going up because there's atail end of all these complex
things. Like, hey, is my, youknow, like, Is my license plate
ready? You know, like, numberplate ready? Like, can I pick it

(20:49):
up? Or is, you know, somequestion around warranty that
only a human would have beenable to answer. Or you have some
question about minor code thatcame up and you don't know what
to do with it, right? Maybe twoyears ago, you couldn't respond
to that, but now you can, andthen you can actually tell them
that this is the diagnosis, youknow, maybe you should come in
and this is how long it wouldtake and how much it would cost.

(21:10):
Maybe you can also add thisadditional service, and hey, I
also see that your car is, like,four years old, and maybe you
want to check out your trade andvaluation, because I just pulled
it out from black book and yournumber, right? Yeah.

And I think it also depends on the amount of
data sources available, right?
If a data source becomesavailable that can tell you
whether or not the license plateis ready, then the AI can ingest
that, right? And more and moredata sources are coming online
every single day, and that's,that's where I think we get that
growing percentage, because it'sall just data, and whether it's
stored in someone's head, aspreadsheet or in an accessible

(21:42):
thing online, like, the moredata that can be consumed, it
can then be shared, right?
Because the natural language is,is, is already there

and, and, you know, like, even, like in the
past, like, you have all theselike data silos, right? They're
locked up in, like, thisinterface and that browser and
this login. But, you know, Idon't know if you've seen any of
those demos where the AI is,like, opening the browser and,
like, you know, log an agent,yeah, human things like as an
agent. So they will makeintegrations even easier. So

(22:18):
imagine it gets easier to accessdata, it gets easier to ingest
it, and it gets very easy tolike, even express it to the
customer, like, through voice.
Like all these three thingshappen at the same time. Just
imagine the possibilities,right? It's crazy. I could pull
anything from anywhere. Andalso, like these llms and these
large language models as youcall them, they know more than

(22:40):
any anybody can know aboutphysics, about chemistry, about
your OEM manuals, about how theengine works. So what are the
odds like we can beat? You know,the model at talking about a
certain problem, because it cansee from all the perspectives
possible to mankind, right? Soput all of it together. It's

(23:01):
going to be a blast. But still,I think the challenge is the
tools always exist. But I alwayslike to tell my team that we're
plumbers, you know? We're takingall these different things and
plumbing it together to make thebest customer experience. What
that experience is going to belike. No one knows, right? That
is the part of the experimentand find out so fun part. Yeah,

well, awesome.
Monique, I it was fun chattingwith you all the way through,
you know, pushing your chips inand soccer and just imagining
what can be with the productthat you're building, that
anybody's building with. AIright now, we can't wait to see
you at ASOTU CON and hang outand learn a little bit more.
Kudos to all you're doing, andthanks for joining us today on
Auto Collabs.

Yeah, thanks for having me. It's going to be a
blast at ASOTU CON. And youknow, we have our racing great
set up over there, like comerace with us.

You, Annie, I don't care if it's 3x 5x 1x if
you're gonna listen to 4000calls, you're dedicated to
building a product that doesn'tmiss

I cannot listen to things at 3x the fact that he
listened to 4000 what happens inyour brain at that point like I
feel like you start to absorbanother layer of conversational
pattern that you probablywouldn't otherwise,

right, like I could probably hear actually a
little bit quicker, is voiceinflection, because listening
fast actually, like you can hearup and down as a pattern across
and so that's actually aninteresting pattern to think
about, because there's so muchwhen it comes to like,
conversational AI that has tounderstand not just what the

(24:47):
words that were said orrequested, but the intent, or,
you know, the the like, what'sgoing on in that person's head
when they say those words andand their level of. Frustration
or energy or joy. You know,golly, it's a complex thing, but
I know that that people like Tomare obviously trying to figure

(25:09):
out, like, how do we how do wefix this? I listen

to three calls, and I'm instantly enraged. So I
don't know,

Unknown (25:18):
I've been there

four car, 4000 calls. I mean, like you're
telling me that is dedicationand also an insane amount of
discipline, because if I listento 4000 calls, I would be rage
eating, I'd be I'd be family.

I mean, why did I just finish this much ice
cream?

One of my favorite things from that conversation
was the soccer metaphor when hishis coach said, you know, it's
not an unsuccessful passes and afailure, it's an unsuccessful
attempt at greatness. That thatreframe was pretty cool, strong,
right? Really? Hey, look, well,you know, I'm not going to go
listen to 4000 calls, but wedon't have to, which is the good

(26:00):
news,

right? You get to listen to Auto Collabs and
get the recap. Even

better. I don't know if there's a recap of the
calls, but regardless, thank youso much to our guests today, and
as always, on behalf of KyleMountsier, Michael Cirillo and
myself, thank you for joining uson Auto Collabs at 1x

Unknown (26:19):
sign up for our free and fun to read daily email for
a free shot of relevant news andautomotive retail media and pop
culture. You can get itnow@asotu.com That's asotu.com
if you love this podcast, pleaseleave us a review and share it
with a friend. Thanks again forlistening. We'll see you next
time youwelcome to an recording. You.

All Episodes

Episode Transcript

Popular Podcasts

On Purpose with Jay Shetty

Crime Junkie

Ridiculous History

.css-15opob5{left:0;position:absolute;top:0.8rem;} All Episodes

.css-14f5ked{margin:0;word-break:break-word;display:-webkit-box;-webkit-box-orient:vertical;box-orient:vertical;-webkit-line-clamp:2;overflow:hidden;}Listening To 4000 AI Phone Calls with Monik Pamecha

Episode Transcript

Popular Podcasts

.css-r6mb8g{margin:0;word-break:break-word;display:-webkit-box;-webkit-box-orient:vertical;box-orient:vertical;-webkit-line-clamp:1;overflow:hidden;}On Purpose with Jay Shetty

Crime Junkie

Ridiculous History

All Episodes

Listening To 4000 AI Phone Calls with Monik Pamecha

On Purpose with Jay Shetty