Yardi CEO Rob Teel on Building Real Estate’s System of Intelligence & How to Serve Clients in the AI Era

Rob Teel is the CEO of Yardi, the property management and real estate investment software company founded in 1984 whose platform serves owners and operators of commercial and residential space worldwide. He joined Yardi in 2003 and spent more than two decades leading its Global Solutions business, overseeing product direction, services, and client relationships across investment management, international, and commercial property owners before being appointed CEO in January 2026. Before Yardi, he was a consulting manager at Accenture focused on ERP and business process outsourcing, where he managed multinational systems implementations for Fortune 100 organizations. Rob is based in Santa Barbara.

(1:05) - Live from Javits Center at CREtech NY
(02:08) Comparing innovation cycles
(03:55) Private company advantage
(06:33) Building AI without chasing short-term revenue
(08:57) SaaS is dead?
(10:38) Leasing chatbot evolution
(14:38) Open vs. closed AI ecosystems
(18:09) Build vs. buy tech
(22:28) Non-real estate tech acquisition: Sidero Labs
(24:10) Yardi lessons year 1 & year 23
(27:47) What’s overhyped and underrated in CRE AI
(31:17) Collaboration superpower: Anant Yardi (Wiki)

🏙️ Learn more:
-Visit Yardi
-Rob Teel on LinkedIn
-Discover Yardi's AI platform Virtuoso

🏙️ In-person Events:
-Finance Forum: NYC | 11/3
-State of Office Forum: NYC | 11/4
-Breakthrough Awards Breakfast | NYC | 12/2
-All upcoming Commercial Observer events

🏙️ Virtual Events:
-AI & Real Estate: You’ve Hired an Alien Superintelligence. Now What?: 11/10

-AI & Innovation Forum: 11/18

🏙️ Connect with Tangent:
-Edward Cohen on LinkedIn
-Zach Aarons on LinkedIn
-Tangent on LinkedIn

Creators and Guests

Host

Edward Cohen

Host & Executive Producer at Tangent

Guest

Rob Teel

CEO of Yardi

What is Tangent: Conversations with Real Estate & Tech Innovators?

A podcast community for real estate and tech leaders, investors, operators and innovators.

A new era is upon us. Tangent💚Proptech unites Proptech founders💡, Real Estate investors and operators🏢, Venture Capital investors, urban leaders🏙 and passionate creators🌎 who are improving our communities and cities. Join us to learn how we can solve present day challenges with innovative technology and greater collaboration.

Join Proptech entrepreneur Edward Cohen, to learn about housing innovation, the future of work and climate tech. We examine these themes through interviews and conversations with Proptech CEO's, real estate owners and operators, and venture investors, and how we can leverage tech and collaboration to help solve the housing crisis, figure out the future of work and how to decarbonize the built environment.

If you would like to get your company’s or organization’s mission featured on Tangent, email us at [e-mail].

Stay curious and always be learning! 🤓

Rob CRE Tech Auphonic
===

[00:00:00] [00:00:01] Hi, I'm [00:00:02] Edward Cohen, and [00:00:03] this is Tangent, a [00:00:04] Commercial Observer [00:00:05] [00:00:06] [00:00:07] [00:00:08] [00:00:09] podcast

[00:00:10] Coming to you [00:00:11] live from CREtech New [00:00:12] York at the Javits [00:00:13] Center, day [00:00:14] one of New York City [00:00:15] Real Estate Week [00:00:16] [00:00:17] Yardi [00:00:18] has been a, the [00:00:19] system of record for [00:00:20] commercial real estate [00:00:21] for over 40 [00:00:22] years. The [00:00:23] question now is how [00:00:24] does that system [00:00:25] evolve from a system [00:00:26] of record to a [00:00:27] system of [00:00:28] intelligence? [00:00:29] [00:00:30] Our guest today [00:00:31] is Rob Teel, [00:00:32] CEO of [00:00:33] Yardi.

[00:00:34] Rob, how is CRE [00:00:35] Tech and New York City [00:00:36] Real Estate Week treating [00:00:37] you? It's [00:00:38] fantastic. Edward, nice to [00:00:39] meet you. Thanks for having me [00:00:40] on the show. [00:00:41] We have to have a debate [00:00:42] on whether it's [00:00:43] C-R-E Tech or CRE [00:00:44] Tech. I've heard it [00:00:45] pronounced both [00:00:46] ways, but we'll go with [00:00:47] CRE Tech. That was [00:00:48] my instinct as [00:00:49] well, so, uh, I think you [00:00:50] nailed it there.

That's [00:00:51] the [00:00:52] hard-hitting analysis [00:00:53] that our listeners [00:00:54] come to hear. [00:00:55] Is it [00:00:56] C-R-E or CRE? [00:00:57] If you ask [00:00:58] me, [00:00:59] CRE Tech [00:01:00] sounds better on the [00:01:01] tongue. maybe [00:01:02] we'll resolve [00:01:03] that issue by the end of [00:01:04] this podcast. But I gotta [00:01:05] tell you, walking [00:01:06] around here, [00:01:07] I've been to CRE Tech [00:01:08] several times [00:01:09] before.

[00:01:10] When the first [00:01:11] wave of PropTech [00:01:12] started back [00:01:13] in, let's call it [00:01:14] 2008, [00:01:15] 2009, [00:01:16] after the [00:01:17] GFC, this [00:01:18] conference started to take off, and [00:01:19] I'm not sure exactly [00:01:20] what year, but [00:01:21] it was known [00:01:22] for [00:01:23] having, you know, the latest [00:01:24] startups, the latest venture [00:01:25] capital-backed [00:01:26] firms. [00:01:27] And [00:01:28] I've seen over the [00:01:29] years those companies [00:01:30] mature and the [00:01:31] conference [00:01:32] mature.

But [00:01:33] when I walk around the [00:01:34] floor today, [00:01:35] it is [00:01:36] crazy. [00:01:37] I, I, I [00:01:38] hardly recognize any [00:01:39] of the booths out there. [00:01:40] They're new [00:01:41] companies, right? There's [00:01:42] something happening at [00:01:43] a very fast [00:01:44] pace. [00:01:45] Obviously, AI [00:01:46] is driving all of [00:01:47] this, but there's an [00:01:48] energy here [00:01:49] that's, uh, [00:01:50] exciting to be a part [00:01:51] of. And I'm, I'm looking [00:01:52] at these [00:01:53] booths as, yeah, [00:01:54] okay, on one hand, they're, [00:01:55] they're competitors, [00:01:56] uh, to some of our [00:01:57] products, [00:01:58] but also, they're the [00:01:59] ones out there [00:02:00] shining a [00:02:01] light on some [00:02:02] new area [00:02:03] of real [00:02:04] estate tech [00:02:05] and AI that [00:02:06] we haven't thought of.

[00:02:07] And so without them, [00:02:08] there's kind of a symbiotic [00:02:09] relationship. For a [00:02:10] large company like [00:02:11] Yardi, with 10,000 [00:02:12] employees and a [00:02:13] 40-year history, [00:02:14] we [00:02:15] need the startup [00:02:16] community to [00:02:17] shine that light. [00:02:18] And in some cases, [00:02:19] you know, we'll [00:02:20] compete, in some cases [00:02:21] we'll partner. Uh, [00:02:22] we have a huge partner [00:02:23] network that we're very [00:02:24] proud of.

But I, [00:02:25] I, I think it's just a, a, [00:02:26] it's an exciting [00:02:27] time [00:02:28] in commercial real [00:02:29] estate right now. [00:02:30] Certainly. And I, I love [00:02:31] hearing how you [00:02:32] [00:02:33] frame the, the [00:02:34] ecosystem that [00:02:35] has been [00:02:36] built to life [00:02:37] here by, by the CRE Tech [00:02:38] team, and they [00:02:39] always do a wonderful job. [00:02:40] But to your point, [00:02:41] you, you joined Yardi [00:02:42] in 2003. [00:02:43] Uh, you've seen [00:02:44] innovation [00:02:45] cycles, [00:02:46] more than one, [00:02:47] through the lens of Yardi.

I [00:02:48] mean, you're... That was [00:02:49] right after the dot-com [00:02:50] bubble burst- Mm-hmm ... right? [00:02:51] Yeah. So in those [00:02:52] 23 years, you've seen [00:02:53] big tech [00:02:54] emerge, uh, [00:02:55] and even go into [00:02:56] some areas of real [00:02:57] estate technology, [00:02:58] especially the house. [00:02:59] But every cycle brings [00:03:00] excitement, it [00:03:01] brings hype, [00:03:02] and [00:03:03] catastrophizing in some [00:03:04] way, shape, or form.[00:03:05]

This cycle, [00:03:06] this AI cycle that [00:03:07] we're in, uh, it, [00:03:08] it feels different. [00:03:09] So how are your [00:03:10] clients and, [00:03:11] and even the Yardi [00:03:12] team [00:03:13] processing this [00:03:14] moment and what you're seeing [00:03:15] on the ground, [00:03:16] how is it different from an [00:03:17] innovation, [00:03:18] from previous [00:03:19] innovation booms that you've seen? [00:03:20] Well, [00:03:21] internet, social, [00:03:22] mobile, [00:03:23] cloud.

You're making me [00:03:24] feel a little old [00:03:25] now. It has been [00:03:26] 20-plus years at [00:03:27] Yardi, and it's been [00:03:28] a, a [00:03:29] fantastic journey. [00:03:30] What I've noticed [00:03:31] about each [00:03:32] period of disruption, [00:03:33] if you wanna call it [00:03:34] that, just on, on a [00:03:35] personal level, you know, [00:03:36] speaking about [00:03:37] Yardi, we [00:03:38] are one of the [00:03:39] largest software [00:03:40] companies in the [00:03:41] world now that's [00:03:42] privately owned, right?

Wow. That [00:03:43] has no [00:03:44] outside [00:03:45] investors, that has no one to [00:03:46] answer to, has no [00:03:47] shareholders, [00:03:48] right? It's a private [00:03:49] company, and, [00:03:50] um, [00:03:51] and in a [00:03:52] period of [00:03:53] disruption like we're in [00:03:54] right now, [00:03:55] it's so important that [00:03:56] no one's asking [00:03:57] you a [00:03:58] question [00:03:59] about, [00:04:00] "Where's your next [00:04:01] dollar of revenue coming [00:04:02] from with [00:04:03] AI? How are you [00:04:04] maximizing profit [00:04:05] with AI?"[00:04:06]

You can really [00:04:07] focus on the [00:04:08] foundational [00:04:09] elements o- of [00:04:10] AI. And so [00:04:11] for a [00:04:12] SaaS company, I, I [00:04:13] spend a lot of time looking [00:04:14] at Salesforce and [00:04:15] HubSpot, and you [00:04:16] could poke your finger at [00:04:17] them and say they've changed their [00:04:18] pricing model five [00:04:19] times in the last two years. [00:04:20] They've changed [00:04:21] their product lineup [00:04:22] five times.

But [00:04:23] we're kind of [00:04:24] similar in that we [00:04:25] are a [00:04:26] large SaaS company [00:04:27] like some of those companies, [00:04:28] Workday and, and, and [00:04:29] others. And so we've [00:04:30] really [00:04:31] monitored what's worked well for [00:04:32] them and tried to [00:04:33] mimic that. One [00:04:34] example might be the [00:04:35] MCP layer [00:04:36] that we include in [00:04:37] our, our primary [00:04:38] product line, Virtuoso [00:04:39] Enterprise.

We [00:04:40] want our clients to [00:04:41] be able [00:04:42] to tap in [00:04:43] through the MCP, the [00:04:44] tools, the [00:04:45] skills that we built, [00:04:46] and it's actually [00:04:47] become a new [00:04:48] part of our product [00:04:49] development [00:04:50] cycle where you [00:04:51] release a [00:04:52] module, let's call it [00:04:53] the [00:04:54] commercial real estate [00:04:55] module within Yardi, [00:04:56] which sits on top of the [00:04:57] accounting system. [00:04:58] But when you release [00:04:59] each module, you [00:05:00] better make sure [00:05:01] the MCP [00:05:02] tools are [00:05:03] updated, that they're [00:05:04] qualified, that the [00:05:05] skills are [00:05:06] good, that we've looked at all [00:05:07] the questions [00:05:08] that, are coming in through [00:05:09] Claude or through [00:05:10] our Virtuoso [00:05:11] Intelligence, and we're, [00:05:12] we're starting [00:05:13] to build this [00:05:14] library [00:05:15] of training on [00:05:16] top of the model [00:05:17] that helps our [00:05:18] clients in the future.

[00:05:19] So it is [00:05:20] a, a period of [00:05:21] disruption, but [00:05:22] in prior [00:05:23] periods, because [00:05:24] we can think [00:05:25] about building the [00:05:26] foundation [00:05:27] and not chase [00:05:28] that next dollar of [00:05:29] revenue with AI, we [00:05:30] have tended to [00:05:31] move and jump [00:05:32] ahead of our, our [00:05:33] peers in periods [00:05:34] like this. [00:05:35] Fascinating. Yeah. [00:05:36] It's the, [00:05:37] the different [00:05:38] incentives, right? By [00:05:39] the s- the structure and, [00:05:40] and who oversees [00:05:41] the company, or who do you [00:05:42] report to, or [00:05:43] what, uh, [00:05:44] timelines you have to [00:05:45] achieve certain [00:05:46] growth.

Yeah. I, I [00:05:47] think that's something that is [00:05:48] underestimated and it's, [00:05:49] like, critical. Yeah Well, I mean, if you're meeting [00:05:50] with a Wall Street [00:05:51] analyst every [00:05:52] quarter, [00:05:53] what question are they gonna [00:05:54] ask? Is, is [00:05:55] AI revenue real? [00:05:56] Right? Is your [00:05:57] profit increasing because of [00:05:58] AI? Are you [00:05:59] reducing [00:06:00] headcount because of, [00:06:01] of AI?[00:06:02]

And a [00:06:03] lot of our [00:06:04] competitors have had to [00:06:05] answer those [00:06:06] questions and make moves [00:06:07] that may not have been in the [00:06:08] best interest of the long run of [00:06:09] the company. And [00:06:10] headcount is, is one [00:06:11] of the ones that we're most [00:06:12] passionate about. We got [00:06:13] 10,500 [00:06:14] employees, and our [00:06:15] growth from a hiring [00:06:16] standpoint has [00:06:17] slowed, which it [00:06:18] should, right?

Because [00:06:19] we're doing five times [00:06:20] the number of GitHub [00:06:21] commits as we did, [00:06:22] you know, two years [00:06:23] ago. The pace of [00:06:24] development is so much [00:06:25] quicker with cloud code. [00:06:26] The ability [00:06:27] for our transaction [00:06:28] teams, our outsource [00:06:29] transaction [00:06:30] teams, some of them are in [00:06:31] Costa Rica, by the way. [00:06:32] Pura vida. They, they can do [00:06:33] their job so much [00:06:34] quicker, right?

So [00:06:35] when I talked to Anant [00:06:36] Yardi about that, [00:06:37] I'm like, "We've never done [00:06:38] a, a layoff in [00:06:39] 40 years. [00:06:40] What if we [00:06:41] build [00:06:42] capacity [00:06:43] here-" Wow. "... and the development [00:06:44] teams and the outsource [00:06:45] processing teams?" [00:06:46] He goes, "If, if [00:06:47] we build capacity, and we [00:06:48] will, the [00:06:49] first thing we do is we give [00:06:50] it back to our [00:06:51] clients." Now, a Wall Street [00:06:52] analyst would tell you [00:06:53] that's not the right [00:06:54] answer, right?

You should do a [00:06:55] layoff. And- Right ... [00:06:56] and we've never done that, [00:06:57] and we, and we don't want to [00:06:58] do that, and we're, [00:06:59] we're very [00:07:00] excited. I'll give you one [00:07:01] example of what we've given back [00:07:02] to our clients. We never had [00:07:03] live support [00:07:04] before. Okay. So the [00:07:05] product, you would enter [00:07:06] a case, we would call [00:07:07] you back at some [00:07:08] schedule, and we would [00:07:09] discuss the [00:07:10] case.

And [00:07:11] most of our primary, [00:07:12] uh, products [00:07:13] now, there's an [00:07:14] option in our [00:07:15] assistant, assistant is like [00:07:16] the little AI bot [00:07:17] that helps you use the [00:07:18] system, [00:07:19] where I can say, "I, I [00:07:20] want a live agent." [00:07:21] Mm-hmm. [00:07:22] And we [00:07:23] could not have done [00:07:24] that three [00:07:25] years ago. It's, [00:07:26] it's AI that's [00:07:27] created that [00:07:28] capacity to give [00:07:29] that capacity back [00:07:30] to our clients.

[00:07:31] And I don't wanna [00:07:32] make this all about, [00:07:33] like, the [00:07:34] disruption period [00:07:35] and how we think about, [00:07:36] you know, building for the [00:07:37] future, but it is [00:07:38] [00:07:39] culturally [00:07:40] a, a really [00:07:41] important [00:07:42] foundation for us. Yeah, it's [00:07:43] unbelievable track [00:07:44] record, 40 years not [00:07:45] having layoffs. And not, [00:07:46] not that there's anything [00:07:47] inherently wrong with layoffs, [00:07:48] but it's not [00:07:49] something that anyone [00:07:50] h- wants to [00:07:51] do.

Yeah. They do it [00:07:52] because they must or [00:07:53] they're pushed to. [00:07:54] And that, that shows that [00:07:55] the human element at [00:07:56] Yardi is something [00:07:57] that you really care about, [00:07:58] and I think that it [00:07:59] translates to how your [00:08:00] clients perceive you [00:08:01] and, and continue to [00:08:02] use you. [00:08:03] Appreciate that. [00:08:04] You [00:08:05] said that SaaS [00:08:06] is dead, [00:08:07] or you've typed [00:08:08] it-

[00:08:09] on the social [00:08:10] medias. That's a bold [00:08:11] claim, coming from the [00:08:12] CEO of one of the [00:08:13] largest SaaS companies [00:08:14] in, in commercial real [00:08:15] estate. So what do you [00:08:16] mean by that? And, and [00:08:17] for a property [00:08:18] manager or asset manager, [00:08:19] fund manager listening [00:08:20] right now, what [00:08:21] replaces [00:08:22] SaaS, uh, and, and how [00:08:23] does that change their [00:08:24] day-to-day?

[00:08:25] Yeah. Well I, I think [00:08:26] it's... I didn't coin [00:08:27] that term. [00:08:28] But I, I think it's [00:08:29] really important [00:08:30] as a company that, [00:08:31] that we consider [00:08:32] that, that we need [00:08:33] to reinvent [00:08:34] ourselves every [00:08:35] day. And I was talking to [00:08:36] somebody earlier [00:08:37] about how [00:08:38] do we think about [00:08:39] competition on this trade [00:08:40] show floor? [00:08:41] I better be [00:08:42] looking at every single one [00:08:43] of my legacy [00:08:44] modules, right?

And [00:08:45] some of the features and [00:08:46] functions have been around [00:08:47] for years and [00:08:48] years. [00:08:49] And our [00:08:50] development team should be [00:08:51] thinking every day [00:08:52] about, do we [00:08:53] continue with that [00:08:54] path or that [00:08:55] screen? Do we need that [00:08:56] screen? Mm. [00:08:57] Can we reinvent that [00:08:58] through [00:08:59] AI? And it's [00:09:00] almost like competing [00:09:01] with yourself. We [00:09:02] have a startup [00:09:03] element to our [00:09:04] culture, [00:09:05] and we [00:09:06] need to [00:09:07] pretend like we're a [00:09:08] startup, and that might [00:09:09] mean that we're [00:09:10] replacing one of our [00:09:11] legacy [00:09:12] modules with [00:09:13] an AI, uh, [00:09:14] native module [00:09:15] And [00:09:16] that's so important.

You could say, [00:09:17] "Well, that might [00:09:18] cannibalize some of your [00:09:19] revenue." And [00:09:20] my view of it is, [00:09:21] if we're not [00:09:22] cannibalizing that revenue, [00:09:23] somebody else is gonna cannibalize [00:09:24] it for us. [00:09:25] Right. Right. And so [00:09:26] we need to think [00:09:27] about... And, and a great example [00:09:28] would be our [00:09:29] AI chatbot. [00:09:30] In, in [00:09:31] multifamily, this is one of the [00:09:32] conversational [00:09:33] AI.

It's [00:09:34] probably one of the [00:09:35] best uses of [00:09:36] conversational AI, which A- [00:09:37] OpenAI is, was [00:09:38] so good at to [00:09:39] begin with. [00:09:40] And we [00:09:41] are on [00:09:42] V3 of that product. [00:09:43] It started [00:09:44] five years [00:09:45] ago [00:09:46] delivering canned [00:09:47] answers through a [00:09:48] chatbot on a website, [00:09:49] right? And it only [00:09:50] did it through the [00:09:51] chatbot. You couldn't call [00:09:52] in, you couldn't [00:09:53] text it.

Mm-hmm. You [00:09:54] couldn't send it an [00:09:55] email. So then we [00:09:56] wired up through [00:09:57] Twilio the different, you know, [00:09:58] channels of [00:09:59] communication, and we rebuilt it [00:10:00] on a different [00:10:01] communication hub. Like, from the [00:10:02] ground up rebuilt [00:10:03] it. The next phase [00:10:04] we had to do was [00:10:05] completely rebuild it. There [00:10:06] can't be canned answers. [00:10:07] It has to be [00:10:08] conversational.

And there's a [00:10:09] nugget of, of [00:10:10] canned answer. And then [00:10:11] you take it even [00:10:12] further and it's like [00:10:13] [00:10:14] V3 now is [00:10:15] completely [00:10:16] conversational, plus [00:10:17] completely [00:10:18] contextually aware. [00:10:19] We never [00:10:20] respond [00:10:21] to a, a chat [00:10:22] question from a prospect [00:10:23] or a resident without [00:10:24] looking at 60 [00:10:25] days of [00:10:26] conversational history. So in [00:10:27] other words, if a [00:10:28] prospect is coming [00:10:29] in [00:10:30] and they're talking to a [00:10:31] leasing agent, it could be [00:10:32] live phone call, [00:10:33] and they mention to the [00:10:34] leasing agent on a live [00:10:35] phone call, [00:10:36] "Uh, this is great.

I'm [00:10:37] really interested. [00:10:38] I'll fill out my [00:10:39] application next [00:10:40] Friday." [00:10:41] Well, the [00:10:42] chatbot is supposed to [00:10:43] follow up [00:10:44] and on a [00:10:45] proactive basis and [00:10:46] reach out and [00:10:47] nudge someone to schedule a [00:10:48] tour, to complete their [00:10:49] application. [00:10:50] When our [00:10:51] chatbot is reaching [00:10:52] out, it's always looking [00:10:53] at this contextual, [00:10:54] you know, en- [00:10:55] environment and all the [00:10:56] conversations that it had.

Whether it [00:10:57] was a previous chat [00:10:58] with a chatbot or whether it was [00:10:59] a live conversation with a [00:11:00] leasing agent. [00:11:01] So in this instance, [00:11:02] it would say [00:11:03] it wouldn't do [00:11:04] anything until next Friday, [00:11:05] and midday next [00:11:06] Friday, it's gonna send a [00:11:07] little text message [00:11:08] saying, "Hey, how's it [00:11:09] going with that application? Do [00:11:10] you need any help?"

[00:11:11] So it's [00:11:12] gone from, you know, [00:11:13] canned answers [00:11:14] to [00:11:15] conversational [00:11:16] to fully [00:11:17] contextually [00:11:18] aware. And one of the things that we [00:11:19] just wired up [00:11:20] in the last [00:11:21] six weeks is the [00:11:22] duplex models. So- [00:11:23] Hmm ... GPT Live, [00:11:24] Gemini Live. [00:11:25] So these [00:11:26] are the [00:11:27] models that now can [00:11:28] listen and [00:11:29] talk at the same [00:11:30] time. Hmm. [00:11:31] Before, when you were [00:11:32] using, let's say you [00:11:33] call into United [00:11:34] Airlines or [00:11:35] Marriott [00:11:36] and, uh, you talk [00:11:37] to their [00:11:38] chatbot, it's very much a [00:11:39] walkie-talkie [00:11:40] type experience.

You've gotta [00:11:41] wait for the chatbot to [00:11:42] finish what they're [00:11:43] saying- Right ... before you [00:11:44] say what you're saying or else [00:11:45] it gets confused. [00:11:46] And we had [00:11:47] some, I, I think it's fair to [00:11:48] say that we've had those [00:11:49] challenges in the past. [00:11:50] But now [00:11:51] with the duplex [00:11:52] models, the [00:11:53] agents can [00:11:54] listen at the same time they're [00:11:55] talking.

Wow. [00:11:56] And, and you [00:11:57] completely bypass that [00:11:58] And [00:11:59] it's made such a [00:12:00] difference. I mean, we had [00:12:01] to just string [00:12:02] together a [00:12:03] number of different [00:12:04] specialty agents to [00:12:05] do what the duplex [00:12:06] models are doing today. So [00:12:07] pretty [00:12:08] exciting. That's super exciting [00:12:09] because it sounds [00:12:10] something that it wouldn't make [00:12:11] such a big difference, right?

[00:12:12] But the pace of [00:12:13] innovation now is so [00:12:14] fast that, like, [00:12:15] things like this [00:12:16] can really move the [00:12:17] needle. I wanna [00:12:18] touch on, I mean, you've [00:12:19] hinted at this a couple of [00:12:20] times when it comes [00:12:21] to working [00:12:22] with, uh, partners, [00:12:23] competitors, [00:12:24] newcomers, [00:12:25] and just new, new [00:12:26] paradigms of [00:12:27] accessing and, [00:12:28] and working with our data [00:12:29] and making decisions.

[00:12:30] So, I mean, [00:12:31] we've had [00:12:32] the likes of, uh, [00:12:33] JP Morgan, [00:12:34] Delta, and Amazon [00:12:35] block [00:12:36] AI [00:12:37] agents from their p- [00:12:38] platform just- Mm-hmm ... in the past [00:12:39] week alone. Mm-hmm. [00:12:40] Realpage did the [00:12:41] same thing with [00:12:42] EliseAI's [00:12:43] latest leasing [00:12:44] AI in, in [00:12:45] multifamily. So [00:12:46] there's a pattern [00:12:47] where certain [00:12:48] incumbent [00:12:49] platforms are [00:12:50] drawing lines on [00:12:51] what agents [00:12:52] can and cannot [00:12:53] do w- inside [00:12:54] their walls or, or [00:12:55] with their data.

[00:12:56] Yardi, you're opening up [00:12:57] MCP access [00:12:58] to Claude, [00:12:59] to [00:13:00] Copilot at the same [00:13:01] time. So [00:13:02] how are you guys thinking [00:13:03] about staying [00:13:04] open [00:13:05] enough to serve [00:13:06] your clients' [00:13:07] evolving [00:13:08] needs while [00:13:09] also protecting [00:13:10] what makes [00:13:11] Yardi, Yardi? [00:13:12] Well, uh, [00:13:13] all of our [00:13:14] conversations over the [00:13:15] last 40 [00:13:16] years [00:13:17] about new [00:13:18] development, about [00:13:19] partnership, about [00:13:20] openness, have been [00:13:21] client-driven, [00:13:22] and [00:13:23] that's the way it's gonna be [00:13:24] right now.

We have too [00:13:25] many large [00:13:26] clients that [00:13:27] have [00:13:28] large relationships with [00:13:29] companies other than [00:13:30] us, and [00:13:31] to the extent [00:13:32] that, that they [00:13:33] need to [00:13:34] partner with other [00:13:35] companies, we need to be [00:13:36] there, uh, for them. [00:13:37] Now, [00:13:38] there is a trend [00:13:39] going on, and JP [00:13:40] Morgan is, is a great [00:13:41] example, right? This [00:13:42] idea of sovereign [00:13:43] AI, [00:13:44] where you keep [00:13:45] things in your house, [00:13:46] and you need to be [00:13:47] protective of that.

Some of our [00:13:48] clients will want [00:13:49] that. Some of them will [00:13:50] want to continue, [00:13:51] uh, to build [00:13:52] things outside of [00:13:53] Yardi, and we've built our [00:13:54] foundational [00:13:55] platform, you know, whether it's the [00:13:56] MCP layer, whether [00:13:57] it's our [00:13:58] APIs, to be [00:13:59] open, and that [00:14:00] position is certainly not [00:14:01] going to change. [00:14:02] With [00:14:03] Virtuoso Intelligence [00:14:04] coming out, one of the [00:14:05] things that- [00:14:06] we've been able to [00:14:07] do is have our [00:14:08] own [00:14:09] self-hosted model, an open [00:14:10] weight model [00:14:11] within the Yardi [00:14:12] cloud.

And you start to [00:14:13] think about [00:14:14] governments and sovereign wealth [00:14:15] and the [00:14:16] desire of so [00:14:17] many of our large clients. [00:14:18] JP Morgan [00:14:19] is a large client of ours. [00:14:20] We haven't talked to them about [00:14:21] this, but if [00:14:22] you want to keep [00:14:23] everything in the [00:14:24] Yardi [00:14:25] bubble, [00:14:26] right? So there's not [00:14:27] a round trip to [00:14:28] Anth- Anthropic. [00:14:29] There's not a round [00:14:30] trip to Copilot.[00:14:31]

We've [00:14:32] structured the foundational [00:14:33] elements to have [00:14:34] self-hosted [00:14:35] models, so these [00:14:36] would be open weight [00:14:37] models. You [00:14:38] train these [00:14:39] models on, [00:14:40] you know, real estate [00:14:41] type of questions. So [00:14:42] that's the, the [00:14:43] post-training [00:14:44] layer of a [00:14:45] harness. Mm-hmm. And [00:14:46] then the most [00:14:47] important thing that we've [00:14:48] done within that [00:14:49] model is, okay, [00:14:50] now everything's within [00:14:51] the Yardi cloud, right?

[00:14:52] You don't have to take this [00:14:53] round trip outside of the [00:14:54] Yardi cloud. Mm-hmm. Now [00:14:55] you've got [00:14:56] a real estate [00:14:57] trained [00:14:58] model. It knows [00:14:59] what, you know, [00:15:00] generically what [00:15:01] occupancy means. It [00:15:02] knows what an [00:15:03] asset manager's [00:15:04] job is and how they like to [00:15:05] view [00:15:06] financials. But we've taken [00:15:07] it a step [00:15:08] further, and we've actually [00:15:09] built client [00:15:10] specific [00:15:11] training.

So it [00:15:12] starts with a vector [00:15:13] database. The vector database [00:15:14] for us isn't, [00:15:15] like, across all [00:15:16] clients. It's specific to [00:15:17] one client, and it [00:15:18] tells, you know, [00:15:19] when AI's trying to [00:15:20] traverse their [00:15:21] data [00:15:22] It's already [00:15:23] indexed and [00:15:24] tokenized all of their [00:15:25] documents. It's already [00:15:26] indexed and tokenized [00:15:27] all of their [00:15:28] objects, like properties and [00:15:29] property lists.

Mm-hmm. [00:15:30] And so I [00:15:31] think the [00:15:32] future of [00:15:33] AI [00:15:34] for a, a, a [00:15:35] large company [00:15:36] and a large SaaS [00:15:37] company has to [00:15:38] consider sovereign AI. [00:15:39] It has to [00:15:40] consider [00:15:41] a post-trained [00:15:42] model [00:15:43] that's smart about [00:15:44] real estate, [00:15:45] but it has to go a [00:15:46] step farther than that. It [00:15:47] has to [00:15:48] consider [00:15:49] the [00:15:50] individual client's [00:15:51] database, [00:15:52] tokenizing as much as it [00:15:53] can to make it very quick to [00:15:54] look up that data, [00:15:55] the [00:15:56] documents.

So building a [00:15:57] harness [00:15:58] around each client's [00:15:59] environment, whether it's [00:16:00] documents, [00:16:01] data. [00:16:02] And then you start to get, [00:16:03] you know, their policies [00:16:04] and procedures baked [00:16:05] into that. And so when someone [00:16:06] asks a [00:16:07] question and asks [00:16:08] for help from a [00:16:09] model, it's [00:16:10] not giving a [00:16:11] generic answer. [00:16:12] It's giving a [00:16:13] company-specific answer.[00:16:14]

This is how we think [00:16:15] about occupancy. [00:16:16] You know, this is how [00:16:17] we weight [00:16:18] occupancy [00:16:19] on move-in, [00:16:20] move-out days. [00:16:21] Like, this is how we [00:16:22] add up our [00:16:23] accounts receivable. [00:16:24] There are things that [00:16:25] have to be [00:16:26] client-specific, and we want our [00:16:27] clients to then go [00:16:28] on... We can't [00:16:29] train [00:16:30] AI and these [00:16:31] models for each [00:16:32] specific client.

[00:16:33] We've built a layer of [00:16:34] client training so [00:16:35] they can add [00:16:36] their rules, their [00:16:37] guardrails, [00:16:38] their training. [00:16:39] We call it [00:16:40] chat guides. [00:16:41] So the chat [00:16:42] guide is the [00:16:43] guardrail around a chat. You can create [00:16:44] a chat guide [00:16:45] for an asset [00:16:46] manager, and it knows the [00:16:47] language of an asset [00:16:48] manager. You can create a chat guide [00:16:49] for the AP clerk, [00:16:50] so it knows [00:16:51] exactly how to [00:16:52] return information [00:16:53] about accounts [00:16:54] payable.

Um, [00:16:55] and these kind of [00:16:56] filters or [00:16:57] pre-prompts are the [00:16:58] training on top of the [00:16:59] harness that makes it [00:17:00] really [00:17:01] sing for an [00:17:02] individual company. [00:17:03] But look, the, the [00:17:04] world's still trying to [00:17:05] figure out how [00:17:06] [00:17:07] open [00:17:08] Is your SaaS [00:17:09] environment. [00:17:10] Our APIs are very [00:17:11] structured and [00:17:12] used by so many [00:17:13] of the companies in [00:17:14] this room.[00:17:15]

The MCP [00:17:16] layer is a [00:17:17] lot more open though, and I [00:17:18] think there's some questions [00:17:19] about how [00:17:20] will companies [00:17:21] want to use that? [00:17:22] How will we [00:17:23] expose all of [00:17:24] it? And I don't have the [00:17:25] answers for that [00:17:26] today, but we'll, [00:17:27] certainly find out. [00:17:28] I, I [00:17:29] generally appreciate your [00:17:30] candidness [00:17:31] and, and your [00:17:32] humility, and also [00:17:33] your confidence in [00:17:34] what you're doing, the [00:17:35] value you provide, [00:17:36] 'cause ultimately that [00:17:37] openness is what [00:17:38] will [00:17:39] allow [00:17:40] clients to continue [00:17:41] feeling [00:17:42] comfortable that [00:17:43] Yardi [00:17:44] has their back, that [00:17:45] Yardi is [00:17:46] evolving with them, [00:17:47] alongside them- [00:17:48] Mm

and is [00:17:49] also letting them [00:17:50] see [00:17:51] how the sausage [00:17:52] is made, if you will, with [00:17:53] the open approach, [00:17:54] while also le- [00:17:55] enabling them to build [00:17:56] whatever they may need- [00:17:57] Well, I mean- ... on [00:17:58] top of you ... it's [00:17:59] not uncommon for me [00:18:00] to walk into [00:18:01] a client meeting [00:18:02] and for them to ask, [00:18:03] "Hey, we're, we're [00:18:04] starting to vibe code some [00:18:05] stuff.

Will I be [00:18:06] able to get access?" [00:18:07] The MCP layer [00:18:08] right now would [00:18:09] be the way to do [00:18:10] that, and [00:18:11] we're doing it ourselves. Like, [00:18:12] we're using [00:18:13] our agent [00:18:14] toolkit and, [00:18:15] and our MCP [00:18:16] layer to create [00:18:17] our agents that are [00:18:18] being delivered to [00:18:19] clients. So [00:18:20] we'll certainly have [00:18:21] that ability [00:18:22] within our [00:18:23] ecosystem, the ability to build [00:18:24] agents.

[00:18:25] But I can tell you, [00:18:26] if the desire's there from the [00:18:27] clients [00:18:28] We're gonna [00:18:29] be there for them. [00:18:30] [00:18:31] So in terms of [00:18:32] Yardi, you [00:18:33] guys have [00:18:34] built, you guys have [00:18:35] acquired [00:18:36] products [00:18:37] across all [00:18:38] functions of commercial [00:18:39] real estate, essentially. [00:18:40] In terms [00:18:41] of how [00:18:42] fast, and, and you [00:18:43] used the D word [00:18:44] before, disruption- [00:18:45] Mm-hmm ... which, uh, you [00:18:46] know, not, not [00:18:47] many in commercial real [00:18:48] estate love that word, [00:18:49] rightfully so, [00:18:50] but us in the tech [00:18:51] world, we love [00:18:52] disrupting stuff for, [00:18:53] for good reason.

[00:18:54] Sure. How do you [00:18:55] think about [00:18:56] staying [00:18:57] ahead of the [00:18:58] curve in terms of [00:18:59] buying versus [00:19:00] building? [00:19:01] You have a team of [00:19:02] over 10,000 [00:19:03] people. [00:19:04] Why wouldn't you be building [00:19:05] everything, right? But at the [00:19:06] same time, [00:19:07] there may be [00:19:08] smaller, leaner, [00:19:09] more focused [00:19:10] teams outside that- Mm ... [00:19:11] may be doing [00:19:12] leaps [00:19:13] and, and that you [00:19:14] may be able to bring them [00:19:15] in-house.

How are you thinking [00:19:16] about your focus [00:19:17] in building in-house versus [00:19:18] acquiring? Well, [00:19:19] first of all, [00:19:20] acquisitions for [00:19:21] us come with a lot [00:19:22] of [00:19:23] overhead, because we, [00:19:24] we never [00:19:25] take an [00:19:26] acquisition and just plug them [00:19:27] in through our APIs. [00:19:28] We take the time [00:19:29] to [00:19:30] rebuild the [00:19:31] software that [00:19:32] the experts at that [00:19:33] company have built.

So [00:19:34] there's almost always [00:19:35] a [00:19:36] rebuild [00:19:37] process, and that [00:19:38] takes time. [00:19:39] And it's [00:19:40] interesting. I [00:19:41] would say the [00:19:42] last year and a half, we've [00:19:43] actually done [00:19:44] fewer [00:19:45] acquisitions than a [00:19:46] typical, you know, year, [00:19:47] two-year period, [00:19:48] and I think the reason [00:19:49] might be that [00:19:50] the [00:19:51] speed of [00:19:52] vibe coding. [00:19:53] I, I feel like [00:19:54] we've got [00:19:55] 10,000 people.[00:19:56]

We've got [00:19:57] 15 [00:19:58] billion square feet of real [00:19:59] estate and 15 [00:20:00] million units of, of [00:20:01] multifamily, [00:20:02] but [00:20:03] the company [00:20:04] doesn't feel like a large [00:20:05] company. It feels like a lot [00:20:06] of startups, [00:20:07] and that's the way we [00:20:08] structure our teams. [00:20:09] There's a focus [00:20:10] area. We pull together [00:20:11] a dev team, a support [00:20:12] team, an implementation [00:20:13] team.

[00:20:14] We're doing it right now in [00:20:15] student, uh, student [00:20:16] housing. We're doing it right [00:20:17] now in manufactured [00:20:18] housing. We're doing it [00:20:19] in [00:20:20] self-storage. Three [00:20:21] new verticals [00:20:22] where we could have [00:20:23] made an [00:20:24] acquisition, but we [00:20:25] chose to vibe code, [00:20:26] [00:20:27] and they've moved [00:20:28] so quickly. [00:20:29] I feel like [00:20:30] they were kind of ac- [00:20:31] Like, they're, they're just [00:20:32] little startups within a [00:20:33] big company.[00:20:34]

But first of [00:20:35] all, Amit Yardi [00:20:36] never had [00:20:37] interest in building a [00:20:38] portfolio company [00:20:39] of companies. [00:20:40] He always [00:20:41] wanted a [00:20:42] holistic [00:20:43] ecosystem [00:20:44] that is, you know, [00:20:45] whether you wanna call it a [00:20:46] single stack or a [00:20:47] single ecosystem, [00:20:48] the, the [00:20:49] technology [00:20:50] is all speaking the [00:20:51] same language, [00:20:52] and there's not ... [00:20:53] You know, we're [00:20:54] not connecting things through [00:20:55] APIs.

We're [00:20:56] building everything [00:20:57] native. Right. Right. [00:20:58] And, and [00:20:59] that is the [00:21:00] DNA of the company. I mean, [00:21:01] Amit Yardi was a, [00:21:02] a, an engineer. He's still [00:21:03] an engineer. He still [00:21:04] has access to cloud [00:21:05] code, which is [00:21:06] terrifying. [00:21:07] But [00:21:08] that [00:21:09] desire to [00:21:10] have a [00:21:11] homogenous- [00:21:12] [00:21:13] technology [00:21:14] foundation [00:21:15] versus [00:21:16] a portfolio [00:21:17] company of [00:21:18] startups or [00:21:19] portfolio company.

[00:21:20] Like, we've always [00:21:21] shied away from that. [00:21:22] So we'll [00:21:23] continue to do [00:21:24] acquisitions. Uh, we did [00:21:25] one [00:21:26] earlier this year. I don't [00:21:27] know if you heard about [00:21:28] Talos. [00:21:29] I have. Has nothing to do with [00:21:30] real estate. I [00:21:31] have now. [00:21:32] Do you want me to tell you [00:21:33] about it? Yes, please. [00:21:34] So [00:21:35] Talos [00:21:36] is an open [00:21:37] source [00:21:38] software that [00:21:39] operates in [00:21:40] the same space as [00:21:41] VMware.

So [00:21:42] if you think about [00:21:43] virtualizing [00:21:44] machines, you start [00:21:45] with bare metal. [00:21:46] Mm-hmm. Uh, a [00:21:47] server, and on [00:21:48] top of that you have to have [00:21:49] an operating [00:21:50] system, which Talos [00:21:51] is an operating [00:21:52] system, and on top of [00:21:53] that, you [00:21:54] virtualize those [00:21:55] machines, right? It's [00:21:56] not- Mm-hmm ... you know, [00:21:57] one machine that you're [00:21:58] dealing with, a physical [00:21:59] machine.

It's been [00:22:00] virtualized, and this has been going [00:22:01] on for years and [00:22:02] with Kubernetes. [00:22:03] Mm-hmm. [00:22:04] And so [00:22:05] on its latest [00:22:06] distraction and [00:22:07] disruption, it has [00:22:08] nothing to do with real [00:22:09] estate. He [00:22:10] was in Amsterdam [00:22:11] last week for a kickoff [00:22:12] launching a new [00:22:13] startup, so it's called [00:22:14] Talos. It's a [00:22:15] virtualization [00:22:16] software with [00:22:17] a director level to [00:22:18] manage all the [00:22:19] virtualization.

[00:22:20] And when we started [00:22:21] showing this to [00:22:22] clients, this idea of [00:22:23] sovereign AI [00:22:24] kept coming up. Like, [00:22:25] yes, [00:22:26] they're struggling with [00:22:27] their [00:22:28] VMware [00:22:29] license and the cost, [00:22:30] and that was kind of an [00:22:31] impetus for some clients [00:22:32] to look around, [00:22:33] but they're also [00:22:34] struggling with, we wanna [00:22:35] host our own [00:22:36] model. And not only [00:22:37] that, we want our own [00:22:38] harness for [00:22:39] our documents, for [00:22:40] our data, for all of [00:22:41] our ERP systems, [00:22:42] non-real estate, [00:22:43] some real [00:22:44] estate, and we want [00:22:45] a [00:22:46] self-hosted version [00:22:47] of Claude.[00:22:48]

And so our [00:22:49] Virtuoso [00:22:50] Intelligence is really kind of [00:22:51] like a, [00:22:52] sometimes we call it baby [00:22:53] Claude. It can [00:22:54] use Anthropic model, it can [00:22:55] use OpenAI, but [00:22:56] it's, it's a chat [00:22:57] session, right? And [00:22:58] with that chat session, you [00:22:59] can build artifacts like [00:23:00] dashboards, [00:23:01] you can retain your [00:23:02] chat history, you can [00:23:03] provision security [00:23:04] so you can see some [00:23:05] data but not [00:23:06] others, which is difficult with [00:23:07] Claude.

Mm-hmm. you can [00:23:08] have drill down [00:23:09] to certain objects [00:23:10] within core [00:23:11] ERP systems. So [00:23:12] this idea [00:23:13] of [00:23:14] virtualization [00:23:15] software [00:23:16] optimized to [00:23:17] manage a [00:23:18] sovereign AI [00:23:19] environment [00:23:20] with an AI [00:23:21] harness [00:23:22] on top of that [00:23:23] to traverse [00:23:24] everything from [00:23:25] documents to data [00:23:26] through a chat [00:23:27] [00:23:28] experience- is [00:23:29] something that we've built for [00:23:30] Yardi.

Uh, we've [00:23:31] built it in our own [00:23:32] ecosystem for our [00:23:33] real estate clients, [00:23:34] and this is the [00:23:35] first example [00:23:36] of us [00:23:37] branching [00:23:38] beyond real estate. [00:23:39] And [00:23:40] I have nothing to do with [00:23:41] this company, by the way. It's [00:23:42] all on Yardi and [00:23:43] his baby. Yeah. [00:23:44] Fascinating. So [00:23:45] it's, we try to keep the, [00:23:46] the company separate and [00:23:47] his startup separate, but [00:23:48] it's, it's a [00:23:49] tremendous [00:23:50] opportunity for him.

[00:23:51] He has so much [00:23:52] energy. Uh, [00:23:53] you wouldn't believe it. [00:23:54] As a founder who's been with the [00:23:55] company for 40 [00:23:56] years- So inspiring ... [00:23:57] he, he couldn't sit still. [00:23:58] So that's what he's up to [00:23:59] these days. So [00:24:00] inspiring. Wonderful. [00:24:01] We'll make sure to check that [00:24:02] out. But [00:24:03] without, uh, repeating [00:24:04] the same number, but [00:24:05] you've been at Yardi for [00:24:06] 23 years, [00:24:07] and you took over as [00:24:08] CEO around 10 months [00:24:09] ago.

First, [00:24:10] congrats on such a [00:24:11] trajectory, and [00:24:12] you still look younger [00:24:13] than me. What's [00:24:14] something that you've [00:24:15] learned in year [00:24:16] 23, [00:24:17] and what's [00:24:18] something that you've learned back [00:24:19] in your first [00:24:20] year when you started? [00:24:21] [00:24:22] Gosh, [00:24:23] uh, the [00:24:24] first year was, [00:24:25] was fun. I [00:24:26] was at, uh, Anderson [00:24:27] Consulting, [00:24:28] Accenture. I [00:24:29] started at Yardi in [00:24:30] 2003, [00:24:31] and [00:24:32] they didn't know where to put [00:24:33] me.

[00:24:34] So [00:24:35] what I learned [00:24:36] is just [00:24:37] pick up the phone and [00:24:38] call a client. I'm not [00:24:39] joking. Like, I [00:24:40] was hired [00:24:41] without [00:24:42] really a title or a [00:24:43] role [00:24:44] and a job- Good problem to have ... and a [00:24:45] job description. [00:24:46] They wanted you. So- They're like, "We [00:24:47] don't know what we [00:24:48] want you for, but we want [00:24:49] you." I walked [00:24:50] in and I, I... When [00:24:51] our new employees [00:24:52] start, I try to, to train [00:24:53] them on this.

Like, [00:24:54] yes, the company has a [00:24:55] responsibility of guiding [00:24:56] your career, [00:24:57] but you also have [00:24:58] a responsibility [00:24:59] of [00:25:00] making your [00:25:01] career. And so [00:25:02] I started chatting [00:25:03] with some folks o- on [00:25:04] the first day, asking [00:25:05] what they were up to, [00:25:06] and, and [00:25:07] they said they were [00:25:08] supporting residential clients. [00:25:09] So I... "What is support?" Yeah. [00:25:10] And it's like, "Well, you get a ticket, [00:25:11] and [00:25:12] you call the client [00:25:13] and see if you can help."

I'm like, [00:25:14] "Oh, that sounds fun. [00:25:15] I'll do that." Yeah. [00:25:16] And, uh, and [00:25:17] so I did. I sat down. [00:25:18] They gave me some tickets. I [00:25:19] started calling clients, [00:25:20] and they'd ask me [00:25:21] questions I didn't know the answer [00:25:22] to, and I'd [00:25:23] say, "I'll get back to you." And [00:25:24] then you go learn, and you [00:25:25] become an expert in one [00:25:26] little thing and then a couple of [00:25:27] things.

And then, [00:25:28] I don't know, what did I [00:25:29] learn over the last [00:25:30] year? Well, at first [00:25:31] I thought, [00:25:32] "Wow, this is... [00:25:33] You know, Amit's [00:25:34] job is not [00:25:35] that difficult." Like, [00:25:36] there's this honeymoon [00:25:37] period of just [00:25:38] going around and [00:25:39] shaking hands and kissing [00:25:40] babies- ... and [00:25:41] visiting clients. [00:25:42] But now [00:25:43] I'm starting to [00:25:44] realize the, the [00:25:45] demands of, [00:25:46] of the [00:25:47] role, but it is a lot [00:25:48] of fun, and I, I [00:25:49] spend [00:25:50] 90% of my time, just [00:25:51] like I did the first [00:25:52] day, talking to [00:25:53] clients.

And you [00:25:54] learn so much [00:25:55] from every [00:25:56] interaction. I try to [00:25:57] get on site with [00:25:58] clients. I, I [00:25:59] wanna hear [00:26:00] what they [00:26:01] want. And if they wanna [00:26:02] integrate with [00:26:03] this company over [00:26:04] there down, you know, [00:26:05] aisle two, [00:26:06] then we're gonna, [00:26:07] we're gonna take a look [00:26:08] at it. If they want us [00:26:09] to build [00:26:10] something on our [00:26:11] product because it's, it's [00:26:12] that homogenous, [00:26:13] you know, single [00:26:14] stack, then we're [00:26:15] probably gonna go for [00:26:16] it.

But it's the [00:26:17] clients that lead us, [00:26:18] and, [00:26:19] and that [00:26:20] hasn't changed in 23 [00:26:21] years. [00:26:22] I think that [00:26:23] homogeneity is, is, [00:26:24] is underrated [00:26:25] oftentimes [00:26:26] because you don't wanna be [00:26:27] chasing or understanding what [00:26:28] a [00:26:29] Frankenstein-like- Well, [00:26:30] imagine- Yeah ... trying to [00:26:31] layer on [00:26:32] AI. Let's say you have [00:26:33] 10-point [00:26:34] solutions or 15 or [00:26:35] 20-point solutions.[00:26:36]

I was talking to a [00:26:37] client or prospect the other day, [00:26:38] had 70 different point [00:26:39] solutions. Wow. [00:26:40] And [00:26:41] the only reason [00:26:42] somebody would move to [00:26:43] Yardi, a, a [00:26:44] large... They're on a, [00:26:45] a decent [00:26:46] platform already, [00:26:47] but [00:26:48] it's to reduce the number of [00:26:49] point solutions. Why would you [00:26:50] wanna reduce the number of [00:26:51] point solutions? [00:26:52] Think about an AI [00:26:53] harness.

Do [00:26:54] you want your AI [00:26:55] harness to have to [00:26:56] learn 70 [00:26:57] different MCP [00:26:58] layers? Or would it [00:26:59] be easier if the [00:27:00] bulk of them, and it's [00:27:01] never gonna be 100% for our [00:27:02] large clients, but let's [00:27:03] say 80, [00:27:04] 85% [00:27:05] of your data comes from the [00:27:06] same place, and [00:27:07] 80, 85% [00:27:08] of your documents and [00:27:09] your workflows are in the [00:27:10] same place. [00:27:11] Now you start layering [00:27:12] Claude on top of [00:27:13] that, and you start [00:27:14] building and vibe [00:27:15] coding on top of that.[00:27:16]

It's like, oh my [00:27:17] gosh, that's so much [00:27:18] easier than trying to figure out [00:27:19] a workflow that [00:27:20] traverses five different [00:27:21] systems. [00:27:22] And so there's a moment in [00:27:23] time right now where I [00:27:24] feel like that [00:27:25] homogeny of [00:27:26] Yardi, that single [00:27:27] stack, that single [00:27:28] connected solution, [00:27:29] is shining [00:27:30] like it's never [00:27:31] shined before. [00:27:32] Because [00:27:33] the things [00:27:34] that you want to [00:27:35] build on AI are [00:27:36] so much [00:27:37] easier if the [00:27:38] transactional layer [00:27:39] and the [00:27:40] workflow is, [00:27:41] is [00:27:42] simple.

Yeah. [00:27:43] Quick [00:27:44] round. What is [00:27:45] over-hyped [00:27:46] and what is [00:27:47] underrated in AI [00:27:48] in real estate right [00:27:49] now? [00:27:50] What is [00:27:51] over-hyped? [00:27:52] I would've said [00:27:53] document abstraction [00:27:54] 'Cause [00:27:55] we've been working on it for [00:27:56] 20 years. We had [00:27:57] machine learning that did [00:27:58] it, and then it actually [00:27:59] kind of worked with [00:28:00] machine learning. [00:28:01] There's still an [00:28:02] element of that.

It is [00:28:03] one of our [00:28:04] premier functions for [00:28:05] commercial real [00:28:06] estate companies is debt [00:28:07] abstraction and, and [00:28:08] lease abstraction. [00:28:09] But when [00:28:10] we built the [00:28:11] product, I [00:28:12] told the team, "I [00:28:13] want it to be valuable [00:28:14] above and beyond [00:28:15] AI." In other [00:28:16] words, I want it to capture [00:28:17] drill down back to the [00:28:18] document and- [00:28:19] Mm-hmm

things that aren't [00:28:20] necessarily... And there's [00:28:21] some structured [00:28:22] queries that go along with [00:28:23] it. So it's not [00:28:24] AI. [00:28:25] It's also not [00:28:26] hype. [00:28:27] It i- it is [00:28:28] something that I would say every [00:28:29] company in this [00:28:30] floor or otherwise that [00:28:31] I talk to that [00:28:32] works with [00:28:33] AI in commercial [00:28:34] real estate, I would say [00:28:35] they always bring up these [00:28:36] functions, so.

Yeah. [00:28:37] I think what's [00:28:38] not hype [00:28:39] is [00:28:40] connecting through [00:28:41] Claude to [00:28:42] your ERP [00:28:43] data, and [00:28:44] particularly in [00:28:45] one role in [00:28:46] every real estate company, [00:28:47] the asset [00:28:48] manager, it, it's like a [00:28:49] secret weapon. When [00:28:50] you're doing an acquisition, when [00:28:51] you're doing a [00:28:52] disposition, when you're [00:28:53] analyzing this property [00:28:54] versus that [00:28:55] property, when you're figuring out [00:28:56] whether I should sell this [00:28:57] property or hold [00:28:58] onto it, [00:28:59] the insights that you [00:29:00] can immediately [00:29:01] get by [00:29:02] connecting [00:29:03] Claude or [00:29:04] Copilot to Yardi [00:29:05] is amazing.[00:29:06]

And now we've built [00:29:07] that in Virtuoso [00:29:08] Intelligence as well, so if you [00:29:09] want a [00:29:10] secured connection [00:29:11] where you're [00:29:12] not, you know, round [00:29:13] tripping to Claude or, [00:29:14] or, or [00:29:15] Copilot, we're gonna be [00:29:16] offering that as [00:29:17] well. But it's the asset [00:29:18] manager role that I think [00:29:19] is not hype. [00:29:20] They are [00:29:21] smarter now. [00:29:22] And so [00:29:23] yes, we can [00:29:24] improve the speed of [00:29:25] transaction processing or [00:29:26] abstracting a [00:29:27] document.

We could reduce [00:29:28] keystrokes. [00:29:29] But you could kinda do [00:29:30] that without AI as [00:29:31] well. I think it's [00:29:32] the insights and the [00:29:33] intelligence [00:29:34] that is [00:29:35] not hype. [00:29:36] Asset [00:29:37] managers, [00:29:38] acquisitions, [00:29:39] dispositions, [00:29:40] executives, and then [00:29:41] any manager in an [00:29:42] organization can be [00:29:43] smarter [00:29:44] immediately about their [00:29:45] team, their data, their [00:29:46] property, their [00:29:47] process by [00:29:48] using AI to be [00:29:49] smarter [00:29:50] versus, versus the [00:29:51] faster side of [00:29:52] things.[00:29:53]

So so- [00:29:54] software engineers are [00:29:55] out and asset [00:29:56] managers are in. Yeah. [00:29:57] Well- [00:29:58] Yeah. [00:29:59] Look, I think it's a [00:30:00] great time to be in [00:30:01] property management. [00:30:02] And the people [00:30:03] component, the [00:30:04] relationships, there's [00:30:05] no way AI [00:30:06] replaces [00:30:07] so many of the [00:30:08] roles within a real [00:30:09] estate [00:30:10] organization. Yeah. I- [00:30:11] we- we're not even gonna [00:30:12] get into, uh, [00:30:13] world models and [00:30:14] robots in this, uh, [00:30:15] interview.

But we'll save [00:30:16] that for a later [00:30:17] conversation. [00:30:18] Rob, [00:30:19] where can [00:30:20] Tangent listeners and [00:30:21] real estate [00:30:22] operators learn more [00:30:23] about Yardi [00:30:24] Virtuoso [00:30:25] and connect with you? [00:30:26] Well, good [00:30:27] question. We just [00:30:28] posted our... YAZ [00:30:29] was last week. That's our [00:30:30] annual conference. [00:30:31] And the general [00:30:32] session is up [00:30:33] online on yardi.com [00:30:34] if you wanna take a look at [00:30:35] that.

It's a good, quick [00:30:36] primer. Our [00:30:37] spotlight sessions for [00:30:38] commercial and [00:30:39] multifamily are [00:30:40] also going up, so, [00:30:41] uh, there's lots of [00:30:42] information out there. [00:30:43] And like I tell [00:30:44] most prospects [00:30:45] and, uh, [00:30:46] clients, it's a small [00:30:47] industry. You wanna [00:30:48] email me or call me? [00:30:49] Go ahead. And, [00:30:50] that's part of the [00:30:51] thing [00:30:52] I love about [00:30:53] this industry, is that we [00:30:54] can know our [00:30:55] clients and we can know [00:30:56] our consumer [00:30:57] versus a true [00:30:58] consumer [00:30:59] e-commerce type of [00:31:00] company.

So [00:31:01] yeah, we'll see [00:31:02] how many calls and [00:31:03] emails I get after that [00:31:04] comment. Or you can come [00:31:05] find Rob at, CREtech [00:31:06] or CREtech. [00:31:07] Last but not [00:31:08] least, [00:31:09] collaboration superpower. [00:31:10] If you could partner [00:31:11] with one person, [00:31:12] historic or [00:31:13] living, real or [00:31:14] fictional, who [00:31:15] would it be? [00:31:16] Oh my [00:31:17] gosh. Such a tough [00:31:18] question. I [00:31:19] mean, this is kinda [00:31:20] lame.

No, no [00:31:21] lame. I mean, if you're gonna say [00:31:22] Elon Musk or Steve [00:31:23] Jobs, that's fine. And, and maybe- [00:31:24] But with- ... maybe kissing up to my [00:31:25] chairman, but- That's okay ... [00:31:26] uh, but [00:31:27] Anant Yardi has been [00:31:28] such a force [00:31:29] in my [00:31:30] career. I've learned so much [00:31:31] from him. [00:31:32] The patience [00:31:33] that he has, and [00:31:34] the one thing that [00:31:35] he's taught me, and I [00:31:36] still work on this, [00:31:37] is [00:31:38] make sure you [00:31:39] listen more than you [00:31:40] talk.

Partnering with [00:31:41] him over the years has [00:31:42] been tremendous, and [00:31:43] he's still giving me [00:31:44] advice. Sometimes I ask [00:31:45] for it, sometimes I [00:31:46] don't ask for it [00:31:47] but- Unsolicited advice, [00:31:48] the best kind. [00:31:49] So that- that's my [00:31:50] answer. I'll go with [00:31:51] Anant. [00:31:52] Rob Teel, [00:31:53] CEO of [00:31:54] Yardi, thank [00:31:55] you so much for [00:31:56] coming to Tangent [00:31:57] today at [00:31:58] CREtech New York.[00:31:59]

I appreciate it. [00:32:00] Great to meet you. Thanks for having [00:32:01] me on the show.

Edward: [00:32:02] [00:32:03] [00:32:04] [00:32:05] [00:32:06] Muchas [00:32:07] gracias. If you learn [00:32:08] something new or [00:32:09] enjoy the [00:32:10] conversation. Text a [00:32:11] friend the link to this [00:32:12] episode right now.

[00:32:13] [00:32:14] You can find [00:32:15] links, [00:32:16] resources, and ways to connect [00:32:17] with us in the show [00:32:18] notes [00:32:19]

[00:32:20]