My favorite is definitely the Leo companions. I think that's really taking Leo to the next level.
Can you share some stats on how AI is helping currently in our business?
When we start a project, we actually define what success looks like before we start working on it.
Welcome back to perspectives unlocked. I am Leo and today we are talking with the three people who shape the future of real everyday. Tamir Poleg, CEO. Pritish Demani CTO. Jenna Rosenblat COO. In this episode, they break down how we are building technology that agents actually love, how we balance speed with excellence, and how AI companions will transform daily agent workflows. You will hear what drives their decisions, what keeps them pushing forward, and how they see the future of real. Let's get into it.
We're announcing some amazing innovation this week. I think that some of the things that we're announcing are the biggest tech announcement that we've made in the past 11 and a half years. So, I'm super excited and I guess that each one of us has their own favorite part of innovation. I know what's my favorite. So maybe we should start by just talking about what your favorite is.
I don't know what your favorite is. What's your favorite?
I think there's an overlap here.
I don't know. And I guess we should also talk about the journey. What made us get to that point? How did we decide to even develop those things? So, maybe we should start with Jenna. What's your favorite?
My favorite is definitely the Leo companions. I think that's really taking Leo to the next level. I know that again thinking through the journey, it's taken some time to get here, but when we think about unlocking the next level of service for agents, whether that be through broker questions or just regular support, it's really going to allow them to have that fast, easy access to the answers they need. It's super super exciting.
So, I think that when we started this journey, this AI journey, we asked ourselves, are agents going to actually use it or are they going to be annoyed by the fact that they're not talking to a human? And I think that AI has improved so dramatically over the past even last the past year. How do you feel about the fact that we're giving our agents an AI interface a non-human even though it's I mean in many cases you cannot distinguish, but did you have concerns about that?
I mean there's always a little hesitation when you're introducing any type of change, but I think one of the things we did well is we're still giving the option to ultimately get to a human if needed, right? And so I think by easing people into it, it's giving them a chance to get comfortable, use it, but still not release them completely from the true support that they know. So it's gone a lot smoother than I probably originally anticipated and I'm just really excited to now see with these companions. I think it's just next level for them.
Can you share some stats on how AI is helping currently in our business?
Yeah, I mean right now about 50% of the time, AI is interjecting in the support flow and answering questions, right? So agents are getting their answers immediately 50% of the time when they're engaging with LEO, which is crazy. They're not having to wait for anybody to respond. And we're just continuing to improve. I mean, I know that you're working a lot with your team, Pesh, to just continually push that needle further, but agents need help when they need it. They don't want to wait. And so by doing that, they're getting everything that they need right when they need it.
So we're improving the level of service at the end of the day.
There's an aspect that people don't talk about very often, but investors do. Investors care tremendously about the efficiency and how AI is going to impact our business. And I think that if we're looking at all of the stats and comparing it to other companies in the field, it seems like we are extremely efficient. Do you think that there is room to become even more efficient with AI?
Absolutely. So we recently built out an AI and automation team within operations that's really helping the entire company with different workflows. And what that has allowed is one has allowed Fesa's team not to have to worry as much about internal operations and really focus on the innovation side to launch things like these Leo companions. But then there's so much opportunity internally just to make us even more efficient than we are today. So we're just getting started. We recently launched this team and it's really, really exciting what they've been able to do in such a short period of time. I think we're going to continue to see us get more and more efficient, allow us to deliver results even faster for our agents and just super exciting.
Can you get us behind the scenes on how those decisions are made? Like how do we think about those ideas? How do they come to life? Who decides who tells the dev team what to do and what's needed? And how do we know that those are things that agents are actually going to use? Because one of the biggest problem is that some brokerages are trying to purchase third party technology but they end up getting I don't know maybe 10% adoption. How do we know that agents are actually going to use our technology?
I think it's probably a great question for you Pesh.
Yeah I mean so I think agents when they start their journey at real they're almost pretty much forced to adopt the technology right like they can't do transactions at real if they're not downloading the app. they even can't apply to join real if they are not part of the application and the kind of agents that we are attracting at the moment are all for the majority of them are technology first they are you know I had an interesting conversation yesterday with somebody who's asking me are you worried that agents are thinking that you're trying to replace them or you worried that agents are thinking oh they're getting replaced by AI and my response was like hey I don't think those kind of agents are joining real in the first place and if there. I mean AI is about to replace like a lot of people in many ways but are they getting replaced or is it additive are they making them superhumans I think agents are about to become super agents and whoever adopts AI is going to become a super agent right so with real adoption is not a if it's a when so when they open the app and when now when they click on Leo all the companions show up I think the question is how much do they believe that the technology is going to work and that's where the last step is is the earning the trust fact. So they'll get an answer from broker companion and then they may or may not trust it. they will escalate to a broker and then the broker will answer the same thing and the broker hopefully the broker says what did Leo say because they came through Leo and actually the broker sees the transcript right so they actually know what Leo said in the first place and just reaffirm it and you get like three reaffirmations and slowly people realize that it's kind of waste of time to call the broker of course there are times when we need brokers which are more difficult answers that require careful thinking analysis and EQ analysis and stuff like that. But basic questions like Leo should be able to answer all conference. I've been going around walking around with my app in my hand and I'm asking agents come up with the most difficult question you can think of. And at the moment it hasn't failed me. And at the conference, I'm going to ask everybody, all 2,000 agents to ask one difficult question as their homework before the end of the day. And then we're going to analyze all 2,000 questions that we get. personally.
Those are real estate related questions.
Real estate related questions that they would ask the broker. Ask Leo. Just make one up,
right? And that will give us 2,000 questions that we can test against in one day.
so take us behind the scenes because you mentioned that moment of launching a product whether it's good enough or not. So when is this moment when you're looking in the mirror and you're saying, 'Okay, this product is ready to launch. It's good enough.' I mean, we know that there's always room for improvement, but there's a phase or a moment in a product's life where you're saying, 'Okay, it's ready.'
It's a great question and most software companies, not only just real estate, like any software company, a lot of them fail at this. And with respect, like they fail at this because software can always be better. So, what's going to happen when we launch? I'm going to get a lot of text messages that give me examples of what doesn't work, right? I will never see the messages of when it worked. I will see some but not nearly close to the messages of when it worked. So when we start a project we actually define what success looks like before we start working on it. So we have a goal in mind and then when we start working towards it we're making sure that we're only solving the problems that we decided we were going to solve and at the end we say okay how close is this? Obviously we're also driven by deadlines. So then if you want to launch everything at rise, which is what we typically try to do, we go on an epic cutting motion, right? Like we start looking at features and say, okay, is this going to make it or not make it? If not make it, is this a do or die feature or is this something we can live with? So even right now there are features that we're pushing right before two days before rise that are do or die features, which is like okay, this is not acceptable, but there are only few left. So my long answer is you define what you want on the onset you track towards it so that you're not shocked when you don't hit it and then you what we call minimum acceptable product as soon as you see it you start using it and then you have to go with it. So it's an MAP.
It's an MAP. Internally we use MAP as because MVP is when you're a true startup where you're very young. You have nothing. You have nothing to lose. You have no customers. You're just saying let's put something out there so we can start seeing traction.
We have a lot to lose. So we can't just say, 'Oh, let me put an MVP out there.'
So we made up this word called minimum acceptable product, which is what's going to be the minimal product that our customers will accept and say, 'Okay, you know what? This is good enough. I'm going to use it and like not throw up on like the first view, right?'
Yeah. So, maybe that's a good segue to what's your favorite announcement of the week?
My favorite announcement of the week is going to be Haleo obviously, but like I mean we've waited someone asked me how long did it take you to build this product and my answer was five and a half years because it takes a lot of thinking and strategizing to this moment. It's not free. Like a lot of stars had to align. AI had to show up like the vision was there but like the technology wasn't right we were two months into AI becoming mainstream when we launched our first AI product two months into it right so we were not behind on this right however between last year when we told everybody hey we're going to give you a phone number where you can text because that was the only thing that was possible and then 6 months after that voice technology took over completely and had zero latency like you can talk to AI voices today and you could get tricked into thinking that it's a human, right? It's very very close. So then zigzag again to get voice first on Haleo. I think I truly believe this is the first time I feel like we've built something that nobody's built before.
But I think that goes back to what you were saying about the acceptable product, right? Like you could have said, 'Okay, we've got these phone numbers. We talked to people about it. Let's just push it out.' But knowing that in your heart you felt like there was something off, like it wasn't ready yet. And so you waited to make sure that okay this is actually going to be a really awesome acceptable product for our agents which is really cool
and again like I truly feel like this is something that nobody's built before.
Yeah I agree and it's been a long wait I think that the market has been anticipating that and asking us questions but the technology kept evolving and I think that the opportunity is so much better right now with voice and video as well.
Yeah I agree. I mean video will be the next frontier. Agentic AI will be like companions in many ways is our foray into agentic AI right so just imagine what will happen next rise right like with this companion concept and the idea of delegating your work to the companions which is like the setup for the whole idea so today you have support and broker which is just imagine the different kind of companions you could have
but you just tell the companion go do something for me not asking questions which is what agentic AI is right imagine like saying hey go run my social strategy for me, which means look at your emails, understand your campaigns from yesterday, look at current world events, make the social art for you, analyze, analyze data when you have the most engagement, whatever it is, and then post it for you, right?
Yeah, that's like true delegation of work. And that's just social like imagine all the different things you could do. So, I think like that's where everything is going. Aentic AI is the keyword of the year, right? Like everything is going in that direction.
Yeah. It throws me back to about two or three years ago when we asked our agents please specify all of the tasks that you're performing for buyers. And I remember that Google sheet and it had like 200 different tasks. And I think that this is how our AI or hey Leo kind of strategy started by trying to identify okay what can we automate and I think that when we looked at that file 80 or 90% of the tasks were automatable by AI. So
but it all starts with just a simple question that we ask our agents at the end of the day just tell us and we take the information we analyze it we think about how can we apply innovation into all of this and then we come up with a product and sometimes people don't see the process between asking that simple question and an agent that send us a five-line email with a few tasks and then two and a half years later there's such a robust and complicated product. That was one.
Yeah. And even Hale Leo like I mean once people start using it I think they're going to realize that it's so organic in what I mean by organic is when someone is talking to Leo a buyer the questions that Leo will ask you is going to be unique and independent because everybody talks to someone differently. So you might say hey I'm looking for a home in San Francisco 2 to3 million. I have a family. You can frontload all the information or you could say I'm looking for home in San Francisco and how you will react after that is a completely independent path which is what true AI is like it's not like remind me when the next birthday is when my customer's birthday is like it's not helpful that's called calendar reminders that's not AI so I'm super excited I think it's going to again listen like disappointment that the product is not good enough is actually a sign that they care, right?
So, and that's how technology see it, right? Like most every time you hear all these noise or news around product launches, you the negative news obviously any kind of negative news just goes faster, right? So, you hear all the negative things. So, when we get negative stuff, we just say, 'Okay, that's our next road map, right?' So, I'm super excited.
Yeah. Yeah, I think that my favorite announcement of the week I feel like Leo Oh, sorry. Wallet is our baby and we said it before. Yeah. So I think that saying something other than wallet would feel like turning my back on our baby. But no, I think that rewards is amazing. Because I do envision a world where or a reality where brokerage is free to agents. Like if we're successful in getting agents to use wallet or entitlement mortgage, brokerage should be free. Yeah. Or at least you could redeem point to that extent that brokerage becomes free. And I think that the point redemption is just a sign of where we're going. And hopefully agents will take advantage of that. So I'm kind of in between points redemption and hey Leo. I think that from a technical perspective, Hal Leo has just the potential of completely transforming agent businesses and just even the journey for consumers. So I think that long-term Haleo is going to be just a beast.
I mean 100%. But what do you want to talk about the Haleo and the payouts for agents?
That's kind of an insane idea.
I think it's an insane idea.
Yeah. Yeah because at the beginning every minute of conversation with Leo cost us money really GPU. Yeah a lot of money
