Latest / NOISE vs SIGNAL by Inman / MoxiWorks CEO Eric Elfman: Building Technology Around the Real Estate Agent

Transcript

MoxiWorks CEO Eric Elfman: Building Technology Around the Real Estate Agent · NOISE vs SIGNAL by Inman

0:00
Today, I am joined by Eric Elfman, the CEO of
0:02
MoxiWorks. And when I posed the question, when
0:05
will the residential real estate market turn
0:07
around and begin to grow again? His answer was
0:11
very, very interesting, a little scary. Stay
0:13
tuned to find out why. Hey everyone, this episode
0:16
is brought to you by Lofty, the agentic AI operating
0:19
system for real estate. Lofty's AI doesn't wait
0:22
to be prompted. It works proactively on behalf
0:25
of your agents to uncover new listing opportunities,
0:28
engage leads, schedule meetings, and drive more
0:32
transactions. Lofty turns every agent into a
0:35
top producer. Visit lofty .com to see how agentic
0:39
AI can put your business growth on autopilot.
0:43
Everybody in the business of home has a pitch.
0:45
Almost nobody gets asked a hard question. This
0:48
show fixes that. Welcome to Noise vs. Signal.
0:52
I'm your host, Inman CEO Tom Bond. Each week
0:55
we sit down with the top producing agents, industry
0:57
CEOs, prop tech founders, and the people shaping
1:01
the future of the business of home. Then we break
1:05
down the week's biggest headlines and decide
1:07
what's noise and what's signal. Let's get into
1:10
it. Good afternoon, everyone. Today, I have Eric
1:16
Elfman with me joining. He is a CEO of Moxi
1:19
Works, the brokerage technology platform behind
1:22
Rise. For this, Eric Elfman built Onnit into
1:26
a $100 million legal tech company. and is now
1:29
rebuilding MoxiWorks from the inside out as
1:32
an AI native platform for real estate. Eric,
1:35
it is so great to have you join me today. Great
1:38
to be here. Thank you for the opportunity. Yeah.
1:39
So, Eric, you and I have something in common,
1:41
and that is we are both what they call in the
1:44
real estate world outsiders. We did not come
1:47
from the real estate world. And I don't know
1:49
if you've felt it, but that's actually a... thing.
1:52
We're kind of looked at sometimes as suspect.
1:54
What could you absolutely know about our world
1:57
that we don't already know better ourselves?
1:59
So first question I have, I just want to talk
2:01
a little bit about your background. Obviously,
2:03
you've grown companies, you've been very successful
2:06
as a leader. Talk to me about how do you translate
2:09
that? from legal tech to real estate tech and
2:12
what that looks like overall. Yeah, look, and
2:14
it's a great point you make. 27 months ago when
2:16
I joined, I was definitely an outsider. But look,
2:19
from my own perspective, then I'll tell you how
2:21
it kind of impacts this business and why our
2:24
financial spots are chosen outsider. Look, everybody
2:27
knows real estate was in such a state of disruption
2:29
that it almost didn't matter the way things used
2:32
to work because everything was new. And for me,
2:36
as a lifelong entrepreneur, I've only founded
2:39
businesses and run businesses before. This was
2:41
new to take over someone else's business. For
2:43
an entrepreneur, disruption equals massive opportunity,
2:47
right? It's room for someone to do something
2:50
different. And I was intrigued by this to prove
2:53
out a couple of theses of mine. One, every good
2:57
entrepreneur and founder has a bit of imposter
2:59
syndrome. Am I a real CEO or do I just call myself
3:03
CEO because I found businesses? So I really did
3:06
want to take over someone's business and I really
3:08
did want to go in a different vertical. I know
3:10
legal tech, but I also had this thesis that building
3:13
software as a service businesses. is the more
3:17
important skill over the domain expertise. Domain
3:21
expertise could be learned, but my 30 years of
3:23
building SaaS businesses, regardless of the end
3:26
user, is what was necessary when you're trying
3:28
to transform a business. And so this was not
3:30
a stable market. This was not a company that
3:33
we needed to just keep doing what we were doing.
3:35
That wouldn't have been a fit for me. Everything
3:37
had to change. The entire real estate industry
3:39
was changing. Our historic customer, the biggest
3:42
enterprise franchisors in the U .S. were getting
3:45
disrupted on some. level and we see that in the
3:47
consolidation today it wasn't do what you were
3:51
doing a little bit better it's change everything
3:53
so as an entrepreneur i'm three things i'm a
3:57
product market fit guy every entrepreneur has
3:59
to spot a problem that they think they know how
4:01
to solve and then i'm marketing and i'm sales
4:04
and when you're inventing something new or transforming
4:06
those are the three things you have to figure
4:08
out very early so it's a perfect fit That's good.
4:11
So one of the things I'm interested to is, you
4:13
know, in the legal world, having been connected
4:15
to that somewhat, you know, billable hours obviously
4:18
is what the whole world's about. AI has not seemed
4:21
to have an appreciable impact on their billable
4:24
hours so far. How does that translate to real
4:27
estate and commissions? Like, do you think AI
4:29
puts pressure on commissions or does it actually
4:32
open up opportunity? I like to say that it raises
4:35
the floor and the ceiling, but I'd love to see
4:37
what you all are hearing and seeing. Yeah, no,
4:39
look, I think that's exactly right. I think AI
4:41
is a huge risk for non -performing agents and
4:44
non -producing agents. It just unfortunately
4:46
is. We don't need as many agents as were in the
4:49
market in 2021 to do so many less home sales
4:53
today. So I think it is a potential disruptor
4:55
to some people's career. But as we see, I think
4:58
it's a force enabler for many. You know, AI should
5:02
replace all of the low value work that isn't
5:04
the value, isn't the judgment, isn't the local
5:07
knowledge. place all of the administrative taxes,
5:10
I like to call it, in any white collar job. There's
5:12
a bunch of stuff you just have to do that does
5:14
not actually pay the bills. AI ought to be able
5:16
to pick up a lot of that. Yes, that means brokerages
5:19
can be more efficient. Yes, that means they can
5:21
do more transactions with less. But I'm a big
5:24
believer that all of this AI is for naught if
5:26
the agent isn't. using the extra time, spending
5:29
more time with their customers, because that's
5:31
what's going to lock in an agent -buyer -seller
5:34
relationship for life, not the use of AI. You
5:37
talked about something that was interesting.
5:39
I was just in Las Vegas at Blueprint, one of
5:41
Inman Group's conferences, and this topic came
5:43
up with a couple of our real estate panels. I
5:46
tried to hold them to a number, and I asked them,
5:48
you know, number of realtors. You know, right
5:50
now they quote about 1 .4 million. Where is that
5:52
going? And then the other one I ask about is,
5:55
you know, housing prices. Where do you see that?
5:57
In general, here's what I heard back, right?
5:59
I heard back that we're probably going to go
6:02
somewhere closer to sub a million real estate
6:05
agents and that housing prices still have about
6:08
20 % to go before they hit bottom. I'd love,
6:11
Eric, to see here what you are seeing. And then
6:13
once we talk about that, let's talk about the
6:15
platform and what it's actually doing. Yeah,
6:17
fair enough. Interesting stats about the number
6:19
of realtors and agents. The reality is it's even
6:22
less than a billion that are actually producing,
6:24
I think, if you look at some of the other studies.
6:26
And so I think it's a great point. Look, there
6:28
are smarter people than me that could talk about
6:30
market dynamics and unaffordability of homes
6:33
and the disconnect between salary increases and
6:36
home increases. Obviously, that's a challenge.
6:39
There's still not enough new homes being built.
6:41
But from my perspective, as a CEO that came into
6:44
this business mid -transformation, We really
6:47
needed to transform the business. The first thing
6:49
I started preaching almost day one is do not
6:52
look for a market rebound or lower mortgage rates
6:56
or more home sales. Absolutely. It will come,
6:59
but it may never come in the form that we have
7:03
seen in peaks before. Just like valuations for
7:05
software businesses in the age of AI is now forever
7:09
capped at a lower number. I honestly believe
7:11
that for the first time ever. And so we've just
7:13
had to spend a lot of our energy figuring out
7:15
how to do what we do for who we do it better,
7:18
not try to count on a market rebound. Because
7:21
I still wouldn't bet on many more than 4 million
7:23
home sales anytime soon. Yeah, $4 million seems
7:28
to be where we're at now with 7 .5 % interest
7:31
rates. Do you see that going up appreciably?
7:33
I don't, and I certainly wouldn't bet on it.
7:36
Look, one of the first things I had to do when
7:38
I joined this business was try and diagnose what
7:40
the problem was. There was a bit of a prevailing
7:42
wisdom in this company and the owners that once
7:45
mortgage rates come below 6%, home sales are
7:49
going to take off again and everything. And I
7:52
had to do a bunch of research to prove that wasn't
7:55
true. The margin pressure in these big brokerages,
7:58
you know, happened a decade plus ago, this compression.
8:01
So even in the best of days in 21, margins were
8:05
low, maybe not grocery store thin low, but they
8:08
were low. And when home sales dropped, it went.
8:12
negative. And so I think the consolidation we're
8:14
seeing is a power play by the few, but it's a
8:17
weakness of the traditional enterprise franchise
8:20
or model, quite honestly, that is a weakness.
8:24
And I think we're seeing the effect of that today.
8:27
So that's interesting. Let's stay there a minute
8:29
because I'd love to because I agree with everything
8:30
you're saying. But what's the path forward beyond
8:33
consolidation? Because it seems like now some
8:36
of the larger consolidators are trying to grab
8:39
and grasp as many bits and points as they can
8:42
in the process through, you know, all these different
8:45
things that have become controversial, whether
8:46
that's private listings or, you know, dual agent
8:49
or selling within or, you know, creating their
8:51
own version of Zillow. And it's across the market.
8:53
Is that also symptomatic of that? What is the
8:56
alternative, do you think, as an outsider looking
8:58
in? Yeah. I mean, look, you actually see two
9:01
different dynamics. It's like a barbell. You've
9:03
got the consolidators and we've got some of them
9:05
as customers that are very successfully consolidating.
9:08
But then you've got these teams of agents that
9:11
has been on the rise for a decade more as well.
9:13
These power teams that are small, efficient,
9:16
that kind of want to remain more independent
9:19
as being consolidated or don't want to become
9:21
part of some of these big tech forward brands.
9:24
So you kind of see both of these things that
9:26
are being successful in my mind. It's the folks
9:29
in the middle that are a little more status quo
9:32
that I think are at the biggest risk. Those are
9:35
the ones getting really emanate up. You know,
9:38
for us, and I don't mean to get into our pitch,
9:41
we're trying to. to offer from a tech perspective
9:44
what the big tech forward brokerages are. And
9:46
we know who they are. I don't have to say that.
9:48
We know who's winning in this market. We're trying
9:50
to offer that same caliber of technology to the
9:53
independent brokerages, because at the end of
9:55
the day, that's what these independent brokerages
9:57
are going to lack most is the access to the scale
10:00
and the tech that some of these bigger brokerages
10:02
have. We want to offer them an alternative, but
10:04
you don't have to just go to one of those. You
10:07
actually teed up my next question. So thank you
10:09
for that. Talk to me about a little bit about
10:11
mock. works, what distinguishes you from the
10:14
others out there? Because there's a lot of players
10:15
in this space. And then the other side of that
10:18
is that some of your work you've cited, like
10:20
agent software adoption stuck around 20 to 25%.
10:23
So the TAM then is huge, right? There's a lot
10:26
there, but there's also a lot of play. And I
10:28
get to see in my position. some of the new entrants
10:31
coming in that are, you know, native and starting.
10:34
So talk to me about, A, what distinguishes you
10:37
all? And then how do you break through in the
10:39
market with all the different players that are
10:41
there now? Yeah, it's a great question. Look,
10:43
we are 15 years old as a company. By the young
10:46
startup standards, we're a legacy vendor and
10:48
all the negatives that that connotes, right?
10:50
But, but... Legacy. 15 years, yeah. Right. Exactly
10:55
the strength of the MoxiWorks business. Where
10:58
did MoxiWorks win? And we were one of the bigs,
11:00
right? One of the three bigs, if you would call
11:02
it that. Where did we win? We won at the biggest
11:04
enterprise franchisors in the U .S. And so we
11:07
really focused. predated me. We focused our building
11:12
of software for those big enterprise franchisors.
11:14
I would argue less for the agents themselves.
11:17
We were serving the brokerage and the reporting
11:19
and a bunch of other stuff first. And I'll tell
11:22
you the secret about why this business needed
11:23
to be transformed. We were selling to people
11:25
that were paying for 100 % of the agents in their
11:28
networks, despite that low adoption. And so what
11:31
is the first thing that happens when home sales
11:33
drop? These brokerages look at why am I paying
11:36
for software that agents won't use? And so first,
11:38
contracts had to be right size and so this business
11:41
had to for the first time start going down market
11:45
into the middle market and start thinking of
11:47
the real buyer of our technologies the agent
11:50
even though we still sell to the brokerages and
11:53
so we had to kind of change everything. And so
11:56
what is it that I saw that brought me here that
11:59
I think is still unique? We are one of the legacy
12:01
vendors, which to me means enterprise grade.
12:04
We can handle the scale of the largest brokerages
12:06
in the United States. Most young companies fail
12:09
to go from really great idea to scale. That's
12:13
a very hard shift. That's why you see a lot of
12:16
vendors get to 1 million, 5 million, even 10
12:18
million, not very many. bump through to 100 million
12:22
at real scale, because it is hard. Which you
12:24
did previously, though. Yeah, 100%. And so we're
12:27
an enterprise grade vendor. We know how to deal
12:29
with the biggest brokerages in the US and really
12:31
Australia, New Zealand, where we are dominating
12:33
as well and coming on strong in the UK. But we
12:36
had the opportunity to write software like a
12:38
startup. Most private equity backed five or six
12:42
year old companies in a portfolio don't get the
12:45
opportunity to rewrite software. But because
12:47
of the state of the market and our customer changing
12:50
so dramatically, we had to rewrite everything
12:53
to have a future. there would be no future for
12:56
the business. And so we're shooting this gap
12:59
in between the enterprise grade legacy vendors
13:01
and the young pups. It's easy to have a great
13:04
idea and a great demo if you don't have 400 ,000
13:07
agents using your software, right? It's really
13:09
easy to show something cool. We're innovating
13:12
like them while being enterprise grade. And so
13:15
what are we today? I would say we are the only
13:19
AI native sales and marketing platform for residential
13:23
real estate agents. There are great point solutions
13:25
that are AI native, but good dumb luck. Every
13:28
entrepreneur needs dumb luck. We started rewriting
13:30
our platform in 24 when AI was becoming so real
13:34
and tangible as a force multiplier for white
13:38
collar workers that we... actually spent six
13:41
extra months and went backwards to make sure
13:43
we were AI native from the ground up. There is
13:46
a difference from AI or agents plugged on to
13:49
the side of your database and being AI native
13:52
from the ground up. So we wind up in this very
13:55
fortunate position where every bit as cool as
13:58
every startup out there, but we've got 400 ,000
14:01
users. You know, it's interesting. Compass just
14:04
talked about releasing their new AI assistant.
14:06
And in the first few weeks, they said about 15
14:09
,000 of 83 ,000 agents used it. That was after
14:12
office tours and rollouts and promos. That's
14:16
very low usage. So I guess there's two parts
14:18
to this question. If the big players can't even
14:21
get their folks to use it and you guys are quoting
14:24
20 to 25 percent, like, is it a software problem?
14:28
Is it a training problem? What's holding the
14:30
bag? Or is it it's going to work its way through
14:33
the big three players and someone's going to
14:35
have to choose them? Yeah. I think it is all
14:38
of that, but some of those are uncontrollables
14:40
in my mind, and I try not to focus on uncontrollable.
14:43
I think it's a software issue, partially, because
14:46
good software that actually helps someone do
14:48
their job better, faster, cheaper, people will
14:51
actually use. You do have this unique nuance
14:54
in real estate and agents being contractors and
14:58
wanting to protect their own contacts from the
15:01
brokerage. So you do have all of these challenges
15:03
in a lot of industries. I never had to deal with
15:05
something like that legally. tech selling to
15:07
corporate lawyers. So it definitely is different,
15:10
but we're just focusing on what we can and having
15:13
to evolve to a world where as these brokerages
15:15
started having such margin compression, a lot
15:19
of their training departments kind of got smaller.
15:22
A lot of their onboarding of their own agents,
15:24
who's going to pick that up? And so we're trying
15:26
to pick up more of the services side of that
15:29
to help people get adoption. But it has to start
15:32
with the right platform and an AI story. that
15:36
is demonstrable very quickly to remove some of
15:39
the tax for the agents. At the end of the day,
15:41
people have to want to use it. Yeah, it's interesting
15:44
because the world we're living in is also getting
15:46
muddied. When you look at these larger players
15:49
like Compass and even Zillow and all of them,
15:52
what business are they in? You could say Compass
15:54
is becoming a technology company. You could say
15:56
Serhant is becoming a technology company. So
15:59
the lines are blending, you know, I think between
16:02
the true, we are simply a brokerage. We help
16:05
people get into home versus, you know, we are
16:07
all things in the business of. selling a home.
16:10
And so do you find yourself competing against
16:13
potentially your own customers? You know, it's
16:15
funny. Why doesn't everybody want to be a software
16:18
company? Look at Tesla's valuation. That's not
16:20
a car company valuation. That's a software valuation.
16:23
So I definitely get it. And look, Keller Williams
16:25
was an early innovator and being so tech forward
16:28
that they had all of their tech stack. And so,
16:30
look, I think it makes sense for folks, but it
16:34
shouldn't work at scale because if you have the
16:37
right vendors in the space, And this could be
16:39
one of the problems, right? If you have the right
16:41
vendors in the space, they should be able to
16:43
scale faster than any individual brokerage can.
16:47
But it simply hasn't been true. And now you've
16:49
got people like Compass with such cash that they
16:52
could pump into the engine that they're doing
16:55
it. But I think even there, you will start to
16:56
see they start to pick up best -in -class software
17:00
vendors to add in to their platform. You know,
17:03
do we compete against them? We don't. But they
17:06
do steal our target addressable market because
17:08
when they buy a brokerage, we can't sell at scale
17:12
to them. I was going to ask you about, yeah.
17:14
So every time they announce another roll up,
17:16
it's an arrow in your heart. Exactly. Shrinking
17:20
Tam. No two ways about it. Look, we have compass
17:23
agents that buy software every day, right? You
17:25
still have this. era of choice for the agents
17:28
and they still will choose their own software
17:30
even if they have to pay extra for it. So we
17:32
still see it. We take those big tech forward
17:34
brokerages as who we want to emulate from the
17:37
scale of what they're doing for their agents
17:39
and help give brokerages an alternative to having
17:42
to be acquired to get access to that level. Yeah.
17:45
Yeah. Very, very interesting. When you all are
17:48
thinking about your customer, is it the team
17:50
leader? Is it the individual agent? Who are you
17:52
approaching in general? Yeah, we are still more
17:54
of a brokerage or a team down as opposed to an
17:57
agent up. I think the further our platform develops,
18:01
the more we'll be able to serve individual agents.
18:04
But we were born in this enterprise scale world.
18:07
Our earliest customers were the biggest brokerage.
18:09
And so we started with this top down kind of
18:12
infrastructure that's a little bit hard to break
18:14
away from. And we still don't want to be in the
18:17
individual. agent sale. But I think at some point
18:21
our technology on our roadmap, it will support
18:24
a pretty frictionless for us. But today it still
18:27
is brokerage, but we're going to a lot smaller
18:29
brokerages, right? We'll see one, two, three,
18:31
five, 10 agent brokerages in our customers today.
18:34
And that, you know, that changes everything when
18:36
you're used to selling multimillion dollar deals
18:38
to selling $10 ,000 deals, different Salesforce,
18:41
different marketing, different implement, everything.
18:44
Yeah, yeah. So what does that look like? Are
18:46
you finding you have to be kind of out there
18:48
more individually at things, meeting people face
18:51
to face? Yeah, that is historically more of an
18:53
enterprise play, right? But we have people at
18:56
Blueprint. We've got people at Riz. We have people
18:58
at your other events. Of course, yeah. to be
19:01
there. The bigger the brokerage, the more you
19:03
have to be shoulder to shoulder. So I think that's
19:05
definitely a part of it. But at scale, the middle
19:08
market and the smaller brokerages more of a marketing
19:11
first play as opposed to a sales and shaking
19:13
hands play. Yeah. Tell me a little bit about
19:15
Rise and what it does and the whole idea of telling
19:19
an agent who to call today and who to reach out
19:21
to. How is that working? How sticky is it? And
19:25
is there an ROI you could put on it? Yeah. Rise
19:27
is so early. We can't point to ROI stats. We're
19:30
desperate for that. We're just new in the market
19:32
this year. But you can start to read the early
19:34
press, right? Century 21 Canada is starting to
19:37
write about it more prolifically. Our early adopters
19:40
are really starting to see the differentiation.
19:42
Look, we had a dozen products when I got here
19:45
in this business. It was too much for a business
19:47
in that state of the market. Some didn't make
19:49
sense. We had back office tech. We doubled down
19:51
on really three core pieces of functionality
19:54
in our stack. It was the CRM out of necessity.
19:57
It's hard to differentiate as a CRM, but it's
20:00
foundational to our email marketing and our CMA,
20:03
our presentation tool. And so that is really
20:05
where we have focused. And the whole design of
20:08
this, and I had moved on into AI. I bought my
20:12
first AI company. in 2020 before the whole chat
20:15
GPT because it wasn't omniscient, but I knew
20:19
things were changing. And so we started to adopt
20:21
it very early. At the end of the day, AI has
20:24
got to remove the friction of doing your job.
20:26
That is what it is best at. One of the earliest
20:29
conferences I spoke at. It was a bunch of brokerages
20:32
talking about how they're using AI to process
20:35
more transactions with less agents, all about
20:37
efficiency. And actually threw my speech away.
20:40
I was supposed to be the technology side of AI.
20:43
I actually threw it away and said, look, all
20:45
of this is awesome. But if you're not using this
20:47
time, if you're just using this to run more transactions,
20:50
you're never going to lock in a buyer or a seller
20:52
for life. Let me tell you what really locks people
20:55
in. And it's the level of service that I just
20:57
experienced when I bought a home, that technology.
21:00
The vendor, the brokerage I used, used Moxi
21:03
Technologies and none of my technology helped
21:06
them keep me as a customer. It was a level of
21:08
service and transaction management that locked
21:11
me in. Yeah. There's so many things I actually
21:13
want to ask you because I'm literally in the
21:14
heart of a lot of these discussions. One of them
21:17
that came up the other day, Eric, was the concern
21:19
around. The fact that most of the companies in
21:23
prop tech who are talking about the use of AI
21:25
are simply utilizing data centers from third
21:29
party as opposed to having their own. And the
21:30
concern about the vulnerabilities around that
21:33
and data ownership and the like. Talk to me a
21:35
little bit about that. Do you guys have plans
21:37
at all to get into that as an overall group to
21:40
have your own data center? Or how are you handling
21:42
that growth of the AI agents that are needed?
21:45
Yeah, and this is where we're a little bit different.
21:47
I'm a believer now that in a software business,
21:50
leadership team, chief product officer might
21:52
be the most important person. And moreover, I'm
21:55
a believer that in 2026 going into 27, your product
21:59
officer ought to be AI native, not AI fluent.
22:02
And so we hired pretty heavy, way heavier than
22:05
we probably should have been allowed to in the
22:07
market, a PhD in computational linguistics who
22:10
was part of the Microsoft Cortana team in the
22:13
early... So she is born and bred AI. And what's
22:18
the point of all of this, right? She's very much
22:20
of an academic perspective. And when you take
22:23
an academic view, you're not just going the path
22:26
of least resistance, throwing out some AI agents.
22:29
You're going from the ground up. How do we protect
22:32
people's data and the end user from potential
22:35
misuse of this technology from the ground up?
22:38
And so we're the only people out there actually
22:40
talking about that type of thing today. It's
22:43
not sexy. It doesn't necessarily sell software.
22:46
And so you won't see us become an LLM. We can't
22:50
compete with those folks. But we have such a
22:54
moat with 15 years of proprietary data. Billions
22:58
and billions and billions of emails, 100 million
23:02
presentations, right? We've got such a mode of
23:05
data that is not publishable to the LLMs. This
23:08
is why LLMs can't just pick up and do everything
23:11
we do. All of these data points feed the AI that
23:15
drives to helping agents be more efficient and
23:18
to wake up and have action items that take a
23:21
simple click for them to actually do their job.
23:25
And so this is where we're focused. Now, look,
23:27
you'll be able to access us through the LLMs
23:29
later this year, but we don't intend to become
23:32
one of them. That's a wasted. Let's talk about
23:35
that. Access you through the LLMs. Zillow is
23:37
making a big deal of the fact that they're now
23:39
embedded into ChatGPT. Are you saying you will
23:42
also be? 100%. Now, if you ask me the next question,
23:45
does it matter? I was going to ask you that question.
23:48
Yes. That I don't know yet, but we were a big
23:53
Claude shop internally. We adopted Claude. Every
23:57
one of us use it through our day. And the value
23:59
of those connectors really is keeping me in Claude
24:03
for a lot of things that I would have jumped
24:05
over to Excel or PowerPoint or email or Teams.
24:09
And so my suspicion is it'll be valuable for
24:13
the end user ages to actually adopt. Claude or
24:16
ChatGPT the way you or I do. And that will be
24:19
the bigger question. What volume of agents will
24:22
use it as an integral part of their day? I don't
24:24
write a spreadsheet anymore, right? I chat through.
24:27
I don't write a presentation anymore. All of
24:30
that is low value work. And so to the degree
24:32
you're that type of person, I think it's going
24:34
to be very valuable. There's not many people
24:36
I would feel comfortable asking this question
24:38
to, but you obviously know what the hell you're
24:40
talking about. So I'm going to ask this. And
24:42
this is just a perspective. And not on the strength
24:45
of LLMs and their power and their transformational
24:48
impact, but on the economics of LLMs and Chad
24:52
and Claude. And it feels to me like a Ponzi scheme
24:55
right now. Let me explain. Because they're continuing
24:59
to lose massive amounts of money. How long can
25:02
that be sustainable? How long they're offering
25:05
us the opportunity to do more and more? And are
25:09
we diving into something that one day the bill
25:11
is going to be due? And none of us are going
25:14
to be able to afford it each year. It's actually
25:16
a story as old as time in tech companies. And
25:19
so I have long made use of young tech companies
25:21
that are buying the market, right, with crazy
25:24
low prices. And you know that the song will change
25:27
at some point. I think the same is true here.
25:29
There's all types of interest conversations around
25:31
these LLMs, the amount of money they're putting
25:34
into data centers and what they're charging,
25:36
right? We have hundreds of thousands of dollars
25:37
in our bill every year now, and we see it skyrocketing.
25:41
scale. And so one topic is you're going to start
25:43
to have halves and half nots. Companies can't
25:46
afford this level of AI for everybody. And so
25:48
some people actually won't be able to get access
25:51
to the tech. What does that do for their career?
25:53
From a cost perspective, there's going to have
25:56
to be a reckoning and it can't continue to scale
25:59
in price to end users as much as it is today.
26:03
I don't know what they do about the billions
26:05
of dollars being spent on these data centers.
26:08
I think someone's just going to have to take
26:09
a bath on that. think these prices can't keep
26:12
going if we're going to have the scale that we
26:14
need. This is unsustainable at this point. You
26:17
know, to me, I'm fascinated by all the monetization
26:20
and economics, whether this becomes a one survives
26:22
and it almost becomes like a utility or, you
26:25
know, it's fascinating to kind of figure some
26:27
of these things out. But that's not for us to
26:29
do today. But here's what I do want to want to
26:32
finish our conversation on. You guys put out
26:34
a white paper and we're going to put the link
26:36
below because I think it's really. fascinating
26:39
and a great piece. I think everyone should read
26:41
it. It's titled AI that earns its place. And
26:44
it really talks about the overall human connection
26:46
and all that. Talk to me a little bit about that
26:48
white paper and give me a specific AI feature
26:51
that you built, tested and pulled because it
26:53
didn't earn its place, if there is one. That
26:56
didn't earn its place. Well, look, let me talk
26:59
about the white paper for just a moment, because
27:00
it's part of our philosophy. We're a vendor that
27:02
talks a lot about customer service, right? I
27:04
used to do this in my almost 30 years in legal
27:06
tech. Lawyers are really customer support to
27:10
the rest of the business. Agents are customer
27:12
service people. It's not doing them a disservice
27:14
to do that. And so AI is an enabler to more customer
27:19
service to me and better customer service that
27:22
has nothing to do with AI. All of us should give
27:24
you more time to do this. Most of the AI that
27:28
we're deploying, I very much like the Apple way
27:31
of doing things as opposed to the Microsoft.
27:34
And I'll tell you my view of that. Microsoft
27:35
puts out a lot of software. It's not always done.
27:38
Apple puts out less, but what is done is great.
27:41
And you always want more. We're trying to do
27:44
more of that path where we're not throwing out
27:46
50 agents. We're trying to be very practical
27:48
based on the data. And so where we're really
27:50
focusing a lot of our energy on today is overnight
27:53
sniffing signals that may mean someone is about
27:57
to go do something, trying to judge intent or
28:00
propensity. Everybody's been talking about in
28:02
this space. I think it's only now starting to
28:04
get real. And so I will tell you what we tried
28:07
to do that maybe was a little bit early in the
28:09
market. We would stitch these signals together
28:12
with actual action. Right. Because you can actually
28:15
have AI, you know, go in and look at the data
28:18
and say, we think this home you've listed is
28:21
five thousand dollars too high, given the days
28:23
on market relative to the industry and that local
28:27
neighborhood. We suggest a five thousand dollar
28:29
decrease. Here's the presentation. We can make
28:32
the change. Right. And so you could actually
28:34
have software do that. I would say we went a
28:37
little bit too far in having the AI actually
28:40
do it and had to back up to. giving people the
28:43
option. So doing all of the work. Here's the
28:46
presentation. Here is the Instagram ad. Here's
28:49
the Facebook ad. Are you ready to go? Click.
28:52
Here's the email. Do you want us to send it?
28:54
And so just that very much at the end of the
28:55
day, a human in the loop. I would call it more
28:58
assistive than autonomous. I don't know that
29:00
this industry will ever want fully autonomous
29:02
AI, and I absolutely get it. We may have stretched
29:06
a little too far on the art of the possible as
29:09
opposed to what's practical. Well, let's finish
29:12
it on that then, because you also have an incredible
29:15
video out, which will also link that is really
29:18
talks about keeping the human connection central.
29:21
Is that built into your corporate DNA now? And
29:24
how does it, is it changing your roadmap? Tell
29:26
me about the video and tell me how it's informing
29:28
your roadmap for your products overall. Yeah,
29:31
it's part of our ethos, right? It is absolutely
29:34
the end of the day about the agent and their
29:36
level of service to their end user because it's
29:38
the only lock -in. Technology is not the lock
29:40
-in. Compass Home, you know, Reels Tech, none
29:43
of that tech is going to lock in a buyer or seller
29:45
to the brokerage. It's that agent, the customer
29:47
service. And so our software has got to give
29:50
people more time and space to do that. And so
29:53
it absolutely is part of everything we're doing.
29:56
Excellent. All right. Eric, I love this chat.
29:59
Really appreciate everything we've talked about.
30:01
I'm going to end on one quick question. This
30:03
is a quick answer, whatever comes to mind. When
30:06
or if, when or will the turnaround for the residential
30:09
real estate industry happen? I'm betting on never,
30:12
although it will come. It's safer to bet on never
30:16
from our perspective. I don't know. What's a
30:20
healthy market? A healthy market is five. It's
30:22
not six or seven, right? If you look long term.
30:25
So when do we get back to five? I bet within
30:27
the next five years, I just wouldn't put money
30:29
on it. Well, five seems reasonable, sustainable
30:32
and historically on mark, right? Yeah, exactly.
30:37
It doesn't feel like 21, but we don't. 21 is
30:40
an anomaly. We can't base everything on 21. And
30:43
unfortunately, I think too many companies built
30:46
their entire future and plans and their investment.
30:49
thesis around 21. If you look at our big customers,
30:53
like the big dollar customers, 21 is when they
30:55
were spending money. Agree to agree. Eric Elfman,
30:58
CEO of Moxi Works, really enjoyed speaking with
31:01
you today. Thank you so much for joining us on
31:03
Noise vs. Signal. Thank you, Tom. Great to see
31:06
you. Thanks for listening to Noise vs. Signal.
31:08
Follow us on Instagram at Inmanews for more conversations
31:13
and industry insights. And be sure to subscribe.
31:16
wherever you get your podcasts so you never miss
31:18
an episode. I'll see you next time.