The Property Stack
  • Home
  • Episodes
    Business 2 Technology 1 Estate Agents 3 AI 3 AI, 1 CRM 1 Property Portal 2 Investment 1 Proptech 2 Data 1 See all categories → All Episodes 5
  • Leave a Review
  • Contact
  • Sponsors
  • About
  • Search
Apple Podcasts Spotify All Players
Follow
Apple Podcasts podcast player icon Apple Podcasts Spotify podcast player icon Spotify All Players
Search
Nov. 4, 2026

Ep. 4 What Listings Don’t Tell You

Ep. 4 What Listings Don’t Tell You
The Property Stack
Ep. 4 What Listings Don’t Tell You
Your browser does not support the audio tag.

What can Europe’s location data teach the rest of the world? GeoSci founder Florian Gruenwald on building a trusted data business from the ground up.

Apple Podcasts podcast player icon Apple Podcasts Apple Podcasts Spotify podcast player icon Spotify Spotify
More
Apple Podcasts podcast player icon Apple Podcasts Apple Podcasts Spotify podcast player icon Spotify Spotify
More
Proptech Data Startup
  • Show Notes
  • Transcript
  • Guest Profile

Why do property listings tell you everything about the kitchen and almost nothing about the location?

In episode 4 of The Property Stack, Hannah Parker and Matt Gayleard (iovox) sit down with Florian Gruenwald, founder of GeoSci, the Hamburg based location data business thats powers neighbourhood insights for property portals and agent websites across Europe.

Florian takes us on a journey of starting GeoSci from his own personal struggle trying to find a quiet flat in Hamburg, to the challenges facing founders in the age of AI and golden learnings of working with cross border data. With Matt and I being data geeks ourselves, we LOVED getting into the reeds on this one.

For more info or to chat to Florian, check out geosci.de

What we get into;

  • Why portals have been slow to add location data, and why AI is changing that
  • The rational vs emotional factors buyers really search for (spoiler alert there are a lot of water and tree mentions)
  • How priorities change across Germany, the Nordics and the Alps
  • A simple three step playbook for portals: show it, make it searchable, personalise it
  • Why “just use OpenStreetMap” is harder than it sounds (boxing schools, anyone?)
  • Build vs buy, why quality and trust still beats speed when people make life decisions on your data

Florian: the answer was it's connectivity. So public transport is super important for people. They want to know about nature. Do you have access to a park, for example? Can you go running in a forest close by? And then it's all about water. So proximity to a river, to a lake. So I think it's interesting. Like on the one side, it's rational factor. So connectivity, how far am I from the next public transport station and how often does the train go? And then all this about nature and water, I think is the side of the emotional side of buying a house where you really fall in love with the property and the environment and you picture yourself there.

 

Hannah: in some of the markets that you've covered where data has actually improved the user experience, improved that lead conversion that somebody operating in a UK market could learn from.

 

Florian: when you've proven that this data is valuable, that your users appreciate it, then you would say, okay, now people can maybe in their own profiles say what they personally care about and then these location components, for example, would adapt to your personal preference. Matt, when you looked for your previous homes that had to have a good school, you would see for every listing what's the score for child-friendliness and maybe have the three best schools shown under that score. So you could directly see does this fit or not fit.

 

Hannah: Today we're very lucky to be joined by Florian Gruenwald the founder of Geosci, a property data business from Hamburg. Florian started GeoSci whilst house hunting himself, and like many of us in this scenario, he realized the listing told him almost nothing about the location. How noisy is it? Is it child friendly? Is it accessible? We have all been there. Today, GeoSci's location data, AI location. Descriptions and interactive heat maps sit inside some of Europe's best property portals and agents-owned websites, helping their home seekers really understand where they're moving to. I'm really happy to welcome Florian to the property stack show today. It's a real data party. I'm Hannah Parker, and presenting alongside me is Matt Gayleard from our sponsor, Iovox..

 

IOVOX SPONSOR: This episode is sponsored by our partner, iovox the data and AI specialists. iovox tracks every lead back to its source and turns every conversation into a booked appointment or closed deal. I've used iovox myself over the years and honestly, they're just one of those no brainer tools that I recommend to nearly every lead generating prop tech business that I work with. For more information, visit iovox.com.

 

Hannah: Hi and welcome Florian to the property stack. Matt and I are very happy to have quote unquote another data geek to join us in the conversations today. Property data is very close to my heart as it's been something I've been involved with for a very long time. So I'm very excited to see where this chat goes. I know Matt is bursting full of different questions as well, and how that could impact. all of the different things from our listeners from what they do from portals, agents and everyone else. So let's just go into it. let let's just understand a bit more about GeoSci, the business, and then I'm gonna find out how you got there. So give us a talk through, elevator pitch of of GeoSci and what it does.

 

Florian: Great. Hi, Han Hi, Matt. I'm super excited to be here. Thank you very much for the invitation. So what Geosci is doing fundamentally, we're helping people find the right place to live or spend a vacation. And we do that by trying to figure out all the important factors that are relevant for people when they make such a choice. from quality of life, schooling, infrastructure, any sort of environmental risks and all of those factors we try to figure out in detail. be able to explain them as well. to do that, because home seekers, for example, are generally on property portals, we very closely collaborate with portals and help them make that data actionable in a way so that they can help home seekers find the most suitable home and also understand why.

 

Hannah: And what's what's got you at this point? So I I read a an interesting stat in the in the the research doc is that you are you might tell me this is totally made up AI rubbish, but you are very hypersensitive to noise. And you you what led you to to found this company was from your own frustrations when trying to find a property, is that right?

 

Matt Gayleard: Right.

 

Florian: Yeah, yeah,

 

Matt Gayleard: Yeah.

 

Florian: yeah, yeah. My wife and I, we moved to Hamburg or we were planning our move to Hamburg. And yes, you're totally right. I don't know where I got this from, but that was not hallucinatory.

 

Hannah: Have my powers.

 

Florian: So I'm quite, I have a very light sleep and I like sleeping. And we, in the process, we're trying to figure out where are actually the homes where we don't, for example, suffer from airplane noise or street noise. and browsing the portals, we did not have that information and it was also pretty hard because either the address was not shown on the property so we could not use it as a starting point for our own research. And even if there was an address, like the official noise maps for example from the Umwelpundesamt, they're not really made for the consumer to browse and to understand. So I thought we're spending so much time trying to find this, trying to figure this out and there must be a better solution and it was I think 2021 or something the pandemic was still there we had a lot of time working from home and I thought okay I'll try if I can build something better and yeah What I focused on first was really only noise and by having built that we were then much better able to figure out noise pollution at least for the homes where we had the address.

 

Matt Gayleard: Yeah, totally.

 

Florian: So this is how we for example found the flat where we are currently.

 

Hannah: So it worked it worked for your search being able to have this this extra information. in your your research that you've had a look at, and both Matt and I spend a lot of time looking at especially portals or different marketplaces and different information that you have, why do you think that there is such a lack of data that sits with property listings, whether it's for rental, whether it's for sale, whether it's an estate agent, a broker's own website, or whether that's on a portal, why do you think that that has historically been such a gap?

 

Florian: Yeah, that's a very good question. I think when I started JOSI, I naively assumed that this gap is something that quickly needs to be closed and everyone in the industry is just waiting for a good solution. But as I found out over the journey, it's not so simple. My theory, right, I guess you both are very deep in the industry, so you probably have an opinion as well, but I'm thinking that originally the value proposition of a portal was for a home seeker to see the entire market. That was such a big improvement compared to originally having to browse newspapers, not knowing if there is maybe any other home that they haven't seen and having it all in more or less in one place. I think this was a huge improvement for home seekers. And then eventually, so that didn't change a lot over the years because portals then got established and the industry got settled. Home seekers were pretty much happy with that experience. And of course, the business model of portals is to get paid mainly by agents. So that was at least for many years the case. So I think there was not so much an incentive of portals to innovate and home seekers were pretty much happy with seeing the market and being able to roughly at least filter for what they need. And if I had known that the point where when I started, maybe I would have been a bit more skeptical about founding GLSI. And I was, I think, very lucky that then after one, two years of founding the company, AI came along and basically reshuffled everything and created this pressure for the entire real estate market to innovate and to offer something new because consumer expectations changed. We're all using TechGPT, Claude. We have very specific questions and we want to get good answers and explainable answers. I think right now this is really moving the market and now creating this opportunity where portals I think are now much more thinking about, for example, including helpful neighborhood information.

 

Hannah: I I have a little like a theory on this, where and and Matt, I I'll I'll come to you in a second because I think you've got good insights on this. I think that a lot of the questions that are being asked by consumers in you know when they're first contacting the portal, the agent through their their lead journey are questions that are this, is this data, that you would hold in this data, you know, important questions to impact their life like how noisy it is, distance to schools, distance to amenities, travel times, etc. And I have I I have a theory that there has been two blockers to this. One that ultimately the agents don't want to put anything on their listing that could be perceived as a negative. So who wants to put this as a noisy area? on a on a listing. Who wants to put something negative on a marketing product, which is a property listing? therefore the agents, the the portal, sorry, are responding to what the agents want to do because ultimately they're paying the bills. And my my second theory on this is also it's hard. Like it people think, there's loads of data that's out there, it's really easy. You can just chuck it on there. I know it's not, you know, Florian, it's not, Matt, you know as well. Like it's not that easy. This data is messy and dirty, and mapping it all to addresses is very complex, which historically has meant very expensive. I think what's switching a little bit now is consumer demand. They're expecting more information and secondly the cost of doing this has now dramatically gone down. Matt, from your insights on the the the person that's that has the most visibility, I guess, between what the customer is asking and the product, where do you do you see a huge amount of data in the questions that are being asked are data points that are just missing from the listings?

 

Matt Gayleard: well I'll ask it backwards. So are these things

 

Hannah: Okay.

 

Matt Gayleard: missing? Yes, because they're currently not in there today. So, you know, we've all at some point wanted to move house and we've worried about this is actually this is the first move I've done in twenty years where I didn't have to worry about a school because my kids are big. So to not have

 

Hannah: Mm-hmm.

 

Matt Gayleard: to worry about being, you know, easy access to a school was was transformative in our search. But up until then, you know, try to be near schools, good schools, all of these things, you know, was was key. And these are probably the big things that people are looking for. And if it's not on the listing, it's probably the next thing they're looking at on when they have ways of doing that. If you move to a new area, that's hard. I think on some of the things that you're looking at, Florin, I think until you've come along I would call this qualitative or quantitative. this house is definitely in this location. It is the X feet from a road. but again, if you were asked to, know, self prescribe your properties and noise level, one you're probably always gonna err on the well I don't think it's that noisy. Tick, it's quiet. so unless you've got some mechanism to measure it, I think you know there's that call and quant kind of part to it. I think with regards to data in what we see in calls, We have the ability to capture all of the information that happens in it. And what happens is we're looking at did the call happen? Did we get the outcomes that the agency was was hoping to get? Was there a valuation, was there a listing, was there a visit, was there a next step? But unless they're actually looking for the underlying information that also comes up in these conversations, what's often happening? You're looking at it at a kind of macro trend. Of the hundred calls, X resulted in the valuation. There's so much data that's hidden in there. I think if someone were to go back into that. same data which you can do with us to say how many people asked about the access to the school the e-charger or the noise that information's there it may be that no one's asked and if you've been doing recall recordings and transcriptions and you've got you know our insights product for example that's something you can easily go back and add so again I think that's where the honest answer is you know I know some of the questions come up often but we only know what people are asking

 

Hannah: use the data to select the data.

 

Matt Gayleard: Exactly. Yeah. What what is what as well as the things that you think you need, what else is coming up most often? And that's data, you know, we run this analysis for people often where we say, Here's what you asked for, but here are the things you didn't know you didn't know. And that's often quite transformative for businesses. what consumers are asking for, but also how teams are maybe answering questions or not answering questions. I think yeah, you don't know what you don't know, I think that's the key thing.

 

Hannah: And Florin,

 

Florian: I think

 

Hannah: you work in Nogo.

 

Florian: that's interesting. Just to connect to Matt's point. Because yesterday I spoke to one of our customers, a big real estate agency, and I asked them about what are people actually looking for in the AI search that we co-created. And there the answer was it's connectivity. So public transport is super important for people. They want to know about nature. Do you have access to a park, for example? Can you go running in a forest close by? And then it's all about water. So proximity to a river, to a lake. So I think it's interesting. Like on the one side, it's rational factor. So connectivity, how far am I from the next public transport station and how often does the train go? And then all this about nature and water, I think is the side of the emotional side of buying a house where you really fall in love with the property and the environment and you picture yourself there. under a tree in the shade in the summer reading a book.

 

Matt Gayleard: Yeah, I and I I think that's a lovely point. I think when when you think about listings currently, they are very rational. But it is as much as it being a very rational decision, it's probably the most emotional one that you're gonna make as well, because it's gonna be your your lifestyle for you and your family for you know, however many years. So yeah, I think the current way in which we serve information is very rational, but it's a very emotional decision. So I think that's a it's it's a really good point.

 

Hannah: Do you see any differences for him? Because I I think you're you're sat in a fairly unique position where you operate across multiple markets. You are in you you collect data and are able to provide that data in many different countries across Europe. Are you seeing different trends of particular data that's interesting or more interesting in some countries versus others?

 

Florian: So first, like our main customers still are in Dach, so Germany, Austria, Switzerland. The Lagecheck, the solution we offer for home seekers to just check locations is live in other countries as well. But I have fewer insights into those. However, what I know from user interviews and from the countries that we are active in, I think in the Nordics it's even more about nature

 

Matt Gayleard: Yeah.

 

Florian: and sea than here in Germany. So that's, I think, one difference. Germany is a lot about public transport in the large cities like Hamburg, Berlin, many people like myself included, for example we don't even have a car so this is super important factor when people choose a new home. In the Alps, so Switzerland, Austria, it's a lot about shade as well. living close to a mountainside where maybe for the entire day the mountain may cover your property in shade. This is an important factor that people want to know about. Yeah, and then of course the regional differences that people who are looking for a home in the north of Germany, for example, they want to know maybe how far is it from the Baltic Sea, from the North Sea. So difference is something people wouldn't ask if you're looking for a home in Bavaria. So kind of if you know the larger context of the region, then in that context want to know really what are the details here that matter for us.

 

Hannah: No. Yeah, I I yeah, I guess that makes sense. I mean even in a specific country you're going to have localized things that impact you more than more than others. are there any things 'cause I think the let's take the UK, I I think that there is still a lack Of this additional data that is provided from real estate-owned marketing tools to the portals. I definitely think that there is a the the turning point is coming very, very soon where consumers are demanding a bit more information. But I think they're probably a little bit behind some of the other European countries. Is there anything, Florin, that you've seen that has worked particularly well? in some of the markets that you've covered in where data has actually improved the user experience, improved that lead conversion that somebody in operating in a UK market could learn from.

 

Florian: Yeah, I would think as a first step just simply making the neighborhood transparent. just simply showing in the listing for the main dimensions that people care about. how well that listing scores. So for nature, public transport, child friendliness and so on. think then there is a choice between is that portal or the company more focused on marketing? Then maybe you would just show the positive factors. Some customers basically use the data like that. Or if you have a stronger focus on the buyer, on the seekers, you can of course show also the negative factors and say, okay, fundamentally that's important as well. And sooner or later it will come out anyway. So why hide it? And I think this is not. a huge project in the end. It's about creating one, for example, additional component in the listing with the factors that for your market matter most. And that, I think, gives you a great starting point to learn about, does this help people make better choices? So do you see any changes in conversion rates and engagements on the listing? And if that is successful, then from that point on, because you already have the data, then it is very easy to make it usable in search. For example, either if you are still in a filter-based world, say a filter for maybe quietness or child-friendly or so, or if you are already using an AI or offering an AI search, then you can make all those location attributes directly acquirable by by an edit-m. So I would do it in that way. Maybe step three then, personalization. when people, when you've proven that this data is valuable, that your users appreciate it, then you would say, okay, now people can maybe in their own profiles say what they personally care about and then these location components, for example, would adapt to your personal preference. So for example, Matt, when you looked for your previous homes that had to have a good school, you would see for every listing what's the score for child-friendliness and maybe have the three best schools shown under that score. So you could directly see does this fit or not fit.

 

Hannah: Personalized searching. And you're and those that that are listening to this thinking, well, yeah, of course, like, you know, this this makes sense. I definitely want to improve the experience. I want to add add more personalization into it. there's there's all this data, it's now, you know, it's easy for me to get. It's never been easier for me to build stuff. what advice would you give to somebody who is thinking, okay, I know I need to provide more information on my listings from a discovery point of view, from user experience, flee quality? But you you I'm sure have got the the scars from from doing this in a sort of pre-AI world and post-AI world. What tips would you share or horror stories would you share back to someone who is thinking about starting this?

 

Florian: Yeah, good question. So yeah, you're right. The data in a way is at your fingertips and with Cloud Code and Codecs, it's never been easier to create software. However, this is quite messy and large in terms of size data to work with, with a lot of complexity. it's... It's very easy to get it wrong in many dimensions. And if you like, can go a little bit more. Should I go a bit more in detail?

 

Hannah: Yeah.

 

Florian: okay. So I mean, for example, many people who would get started, they would say, okay, we have OpenStreetMap. That is

 

Matt Gayleard: And that's

 

Florian: generally great quality data. We just use that. Now, then let's say you want to show the best schools, just simply the nearby schools for each listing. If you want to do that, then you would have to get the data. You can't just use the public OpenStreetMap API because that's not made for this use case. So you would have to host it yourself. For Germany, Austria, Switzerland, this market, for example, this is already 200

 

Matt Gayleard: Market for example a

 

Florian: gigabytes of data. you need to have a quite powerful database that needs to be running somewhere. You need to think about updating this type of data. It needs to be in sync. And then in addition, schools in the OpenStreetMap taxonomy, people are very flexible. anyone can say that something is a school that could be, for example, a boxing school. It could be a rowing school that in the end is tagged in a way that if you just query for schools, you will get all those schools. And maybe you at first won't know that you have a problem because maybe these are This only happens in a few percent of cases, but still a few percent of your listings showing clearly erroneous information. I think this is a problem from a trust perspective. So like all of these complexities of the data are hard to deal with. So I think you would need someone who... really goes in into these details. And then that's question, okay, do you have such a person? And from a P &L perspective, is it for you a good investment to... to pay someone to do that. I think this in the end, maybe it depends a bit on maybe the size of the portal or how strategic or how important they consider having that capability in-house. But I think this is the fundamental choice. It's not an easy problem to solve. You can solve it, but it only makes sense if you really think there's a business case behind it, I think.

 

Hannah: Yeah, but I I can f I I feel the pain 'cause I I I have been there and done that and and seen some of the the scars from from struggling with with with mapping together different different data sets and stuff. It's not as easy as it as it first looks. but I I definitely think that this is something that every one marketing property, from an indiv independent agent all the way through to to a portal, needs to be having like this needs To be a very, very significant moment of their roadmap that they need to be thinking about how to how to bring it in. But in terms of the lead journeys off of the back of that, because the more information that's there can lead to sometimes more questions. users are going to want to have responses to those questions. Maybe they maybe their the data has given them an idea, but they still have questions that come back off the back of it. Like you said, okay, it's new. near a school, but what actually is this type of school? maybe it's to do with religion, maybe it's to do with boxing or education, maybe it's to do with with something else. In terms of the lead journey, where do you think that evolution might end up going?

 

Florian: mean who in the end will end up answering those questions.

 

Hannah: Yes.

 

Florian: I think there are two models probably. One, for example, how also ImusCode in Germany is doing it, that they offer an AI assistant whom you can ask all of those things and

 

Matt Gayleard: Awesome, big two.

 

Florian: hopefully get back an answer that's clarifying your doubts, your questions. I think in the end that will work only to a certain extent. because these AI assistants in the end also only can talk about or should only talk about what they can substantiate with data and not all of that data is really available to them. And I think the question is how can you make them actually helpful without them hallucinating, without telling people what is actually not true, not substantiated. So still that's one model. And I think the other model is to say, okay, eventually there is the switch from the home seeker who will at some point talk to an agent and that agent will own that relation, the personal relationship who will then answer all the questions and can maybe offer their personal touch or maybe someone who knows the area well

 

Matt Gayleard: which?

 

Florian: who can answer questions differently than an AI could.

 

Hannah: Yeah, definitely. It's finding the the the moment that it switches over to the to the the pro, not let the robot do everything that's that's there. I'm sh what do I think?

 

Florian: What do you think? What do you two think? Will the portals all in the future have an AI assistant who's sort of a bias agent? Maybe if you want to say it like that.

 

Hannah: I think that users' expectations want information faster. I don't think patience of consumers is getting larger, it's just getting smaller. So their demand of information and response back is just accelerating. So the company that is able, the product that is able to give them an answer as fast as possible, I think will be the one that wins the next level of engagement, which we hope is then the lead conversion. If a user has to leave the environment they're in, website, app, whatever it is, to have to go somewhere else to answer that question, then the opportunity for them to then come back to convert from leads is significantly lower. So I think the person who is able to answer the most amount of questions fastest will will ultimately win the race You've both... You both work with similar customers, right? Real estate agents and portals. Like they are a a big part, or marketplaces, they're a they're a big part of both of your customer bases. So I am sure that both of you have got to experience a lot of new products that have been created off the back of new data, like data-led products that have come about. What have and I'll go to Matt first, any products or features that you've seen recently across the across the board of your customers that have started to do great things with data that you've thought, this is really smart, like this is gonna this is definitely gonna improve the experience.

 

Matt Gayleard: I I think on the data side, what I'm not seeing the wholesale improvement of how we use data better than I have so much AI agents. So if I was to say what I'm just naturally hearing come up in conversations. Now we sell AI agents, so we often start the conversation, but I think if I go back to November last year, so almost a year, We had a lot of conversations with people about what is your AI strategy, how y you've probably got pressure from your board, pressure from your management team, how you're gonna incorporate AI and improve your profit, your customer experience, whatever. Now we're actually talking to people about the actual things that they're doing with AI and and it feels like it's come a really, really long way in a year. and if you think about how the models have evolved, you know, what Claude could do a year ago to what Claude can do now is is kind of frightening, the difference. So I think what's naturally bubbling up is people recognise that there is a missed opportunity at the moment that calls aren't being answered. So they're looking for AI agents to be able to qualify things and Do things that are currently maybe just being missed. There there is the problem exists, this solves the problem. We didn't create a problem that AI could solve. That's the important thing. I think on the data side of things, I think a lot of people don't know that this data is accessible. and it is possible to have a better understanding of what's going on in your business if you know that these data points exist. And I and I think Different businesses have got a different view of that. They might just be looking at it from the marketing funnel and understanding which channels are performing better than others, what ones have better conversion scores than others. I don't think many are looking at it so much around how can we help the teams deliver a better customer experience. What coaching can we give there? Not like I've seen in other industries where they've really deployed that kind of insight against sales teams, you know, and sales coaching. I I've seen in other environments that's whether demonstrably salespeople, I think that's that's

 

Hannah: Mm.

 

Matt Gayleard: I I I think it's an interesting one. I I think it's I think data's there but it's it's a mixed bag of different businesses, different levels of sophistication, what they have in their tech stack. A long answer I think. But I think I think it's the it's the AI agents that seems the lowest hanging fruit where there is a legitimate problem that they'll see a great benefit from solving.

 

Hannah: Hmm. It yeah, it makes sense, I guess, as the first step into into this world. And and Flo, what about you? Like have you seen some impressive or horror story features that you that have that have come out of the ground recently?

 

Florian: I think I like Matt's point about solving a legitimate problem because since... producing software has become so easy. I think there are many not legitimate problems being on things that are not really problems being solved simply because there is a pressure to maybe use AI and show your investors that you're doing something with LLMs. And honestly, like, Right now we are just in the process of buying a house. The offers from portals or anything from the real estate side where I was impressed by

 

Matt Gayleard: Why?

 

Florian: anything... AI related, there's nothing that comes to mind that really would have made a difference for us. For us, mean, after we have received the complete listing, including the address and so on, then we're very deeply involved with Claude or JetGPT to do a due diligence on the property itself. I mean, that has been super helpful, but that is of portal that's already later. Yeah.

 

Hannah: that's something you're doing yourself rather than something that's been provided, yeah.

 

Florian: Yes, yes, yes, yes. And there's, mean,

 

Matt Gayleard: and that's coming.

 

Florian: if I think about cool features from portals. They still, maybe you know, fin.Norway, the portal from, I think it's Norway. They have a great location information section that is available in almost every listing. Almost all of the agents actually pay for that, which includes all the important factors that people want to know from transportation over socio-demographic to schools and so on. It includes like original voices from neighborhood where people say, hey, this is very

 

Hannah: reviews.

 

Florian: friendly. Yes, exactly. Like all of those, this type of information is there. It has nothing to do with AI. It's been there for a couple of years. It hasn't really changed a lot because it's, think, still as helpful now as it used to be in 2022 or so. I once interviewed Jorg Enheles, the then CPO of the company on that product. And yeah, I still love it. It's still something I've find super helpful.

 

Hannah: Yeah, aer area guides are always possible. It's it's funny what you said about, the the the rush of products to come out. I I've had quite a few conversations with people recently where they're just like, We've burnt through our our entire roadmap, we've just burnt through all, but actually there's a reason why some of these things were just on the roadmap at the bottom of the list and never came to fruition before, because when you have to prioritize what you're gonna build, you have to try and look at the stuff that's gonna add the biggest amount of value. When you don't your kind of matrix scoring is out the window because the cost of delivering something is now Almost negligible, you can release a lot of stuff that doesn't add a lot of value because the the risk cost is really low. That doesn't mean you're actually changing your business. You doesn't mean you're moving ahead with your with the with the business. So it's interesting, it's an interesting question I find to a lot of different people because I think I'm yet to have something that

 

Matt Gayleard: She's like.

 

Hannah: anyone's come back to me and said this product that wouldn't have existed two years ago is significant.

 

Matt Gayleard: Yeah, I still could definitely come and turn.

 

Hannah: significantly changing the value and the experience for someone buying, renting or selling a home. So I'm disappointed but not surprised.

 

Florian: Apart from the pure LLMs, Yeah, I think apart from the pure LLMs, no? Like what you can do with... code or with with ChatGPT. mean either building some small tools that can help yourself in a personalized way. That's what I've been doing in our home search. Or simply like helping us to analyze the property or any important informations around renovations. How much that would cost? What would the timeline be? I think this for us has been a game changer. But yeah, it's not coming from the real estate space.

 

Hannah: Time will tell, I think, where their next investment will will take them. And in your own your own kind of learning journey, Florin, for AI, like you you're talking about using AI in your own business and and moving forward with development. Where's how's your kind of learning journey come from your own background within technology and data? And creating a business and now incorporating AI, what's that learning curve been like for you?

 

Florian: I I started pre-AI basically. I was not originally an engineer. had just had done a lot of courses in my free time and then I was by myself. So I had to learn all the skills, front-end development, back-end development to be able to build the product. But yeah, now of course. We are very heavily using AI to generate code and the learning is from my perspective that there is not yet a clear good or best practice of how to use. AI, at least when it is a critical component where the quality of the results really matter. I don't think we are at a stage yet where we can just say, here, agents swarm, here's the product spec. Now you just go build it, then we deploy it and everyone's happy. That's not working. We see that because we still look at the code and we keep seeing critical errors. so that the AI has done because they maybe misunderstood us or they misunderstood the context or any requirements that were actually not there. So I think the learning is that we, the engineers who have that very technical perspective, myself who maybe lives a bit more in the future, I try to think, okay, how does the world look like in half a year, a year? How do we get there? That... that perspective, we try to put it together and think about where are we right now? What does it mean? And for us, means still having a lot of discussions around features, where is the value being created? How does architecture should look like? yeah, as I said, we're still looking at the code and doing manual code reviews. But I think in a couple of months, this may look completely different. And the important thing is that we still we don't get ahead of ourselves, that we maintain that conversation and try to see okay where are we actually at at this very moment.

 

Hannah: Do you feel pressure as a founder in 2026 with so many new businesses being created in seconds, it feels like, with the cost and accessibility of building coming down so much? Has that changed this year for you as a founder being in that race?

 

Florian: I feel more pressure, I think, to stay a good option in the build versus buy decision that happens internally of partner companies, because the engineering teams say, okay, why don't we just do that ourselves? I mean, all the data is there, I know what we discussed before. And I think there it's important that we keep proving the value and the speed advantage, of course, that they can get. Other startups, mean in Hamburg, for example, there is a meetup that I go to from time to time. They're called AI beavers. They're basically wipe coders who are building stuff and thinking of founding companies. But somehow very often they are not really successful in founding a company because in the end it is so much more complex than just building a product, especially when you're only white coding the product and you don't really fundamentally understand the domain yourself and don't really have the depth to then articulate what you've built maybe and the limitations and so on. So I think the speed advantage that startups have with product building doesn't mean that these in the end are threatening competitors because very often they're lacking things to be a serious business in my perspective.

 

Hannah: Yeah, we'll they we'll we'll see the where they fall out, I guess, over time. And Matt, I would say, you know, IVox is a mature startup. I i it's a startup plus plus maybe in in that environment. But do you and the team a teenage

 

Matt Gayleard: We're we're teenage we're a teenage startup, I think.

 

Hannah: startup, I like that.

 

Florian: FFFF

 

Hannah: But do you and the team still feel that pressure this year, like a a change of of the environment?

 

Matt Gayleard: Yeah, I d I don't think this year feels like any year I've ever had. And I don't think it probably feels the same. You it's just me personally but also I think I've also I think the speed of change, the speed of personal empowerment, capability, you know, i if I go back six months and in the six months before that and the six months before that. Each one is so is transformationally different. So we were using ChatGPT again, just in like I'll pick the commercial team. We were using Chat GPT to help us maybe do a bit of desktop research, maybe send some email, you know, not send emails but draft emails and some complicated things. Where we are now It it's a million miles away. We've we've built our own in-house CRM, our own communication system, we're able to slice and dice data and understand what's going on instantly in ways that would have taken someone in my team nine months ago a couple of days who are good at spreadsheets to do. Whereas we can just do it and it's instant. So I think the kind of instant access to a deeper understanding of what's going on in your business, to then be able to move faster and deploy and build product quicker better it it's night and day from where we were before. but what that means though is that as an individual and as a business you've got to have you've got to make time to invest in using the tools, experimenting with the tools, go down a few creative cul-de-sacs to be able to come back and go, okay well that worked, that didn't work. And it but also I think I think what it's breeding is that as workers we need to develop higher degrees of acceptance to work with ambiguity because things are moving faster. They are they are less fixed than they were. So I think that you know, comfortable with ambiguity, comfortable with learning fast on the job. But still that level of discernment to go, okay, but now when we execute, I've got to have an attention to detail. And I think that they're yin and yang from each other in some ways. high levels of ambiguity, yeah, we'll figure it out as we go. But now it's at that point of deployment, I've got to be really thorough and make sure that it works exactly how it should work. It's perfect and it's tested and all this thing. So I I think we're just crushing that pipeline of, you know, what used to take a year to build a product. You're smooshing it down into a sh shorter period of time, but the quality at each point of those gates, the thinking, the development, the codes, the execution, has to be as good as it was before. So I th so I think this it's accelerating it's it's a completely new way of working, I think, which some people are loving, I think some people are finding hard. Would be my my

 

Hannah: That's a it's a good summary, yeah, definitely. I I think the l the learnings from that I'm taking from both of you are that AI will accelerate people that are already good at this and just it the the people that aren't maybe haven't got the the same level of experience or domain knowledge will you know fall over at some point. and trusting, you know, if you're if you're looking at these new companies as a supplier, I I would just say, you know, be careful. careful, especially when it comes to things like data, because that in the I think there was something that you touched on at the beginning of our of our chat, Florian, which is about trust. You know, if you if you make mistakes, then you'll quickly lose trust. So if you're dealing with a product with that manages data, if you're giving incorrect information, it's over. It's game over. So definitely definitely use people that are experts in this game. I I think that the product

 

Florian: Yeah, actually in our case we are rather... If we make the trade-off between speed and quality, since, as you said, this is such important data where people base important decisions on, we are rather a little bit slower and make sure that what we do is thorough and substantiated. And even if that means foregoing some of the speed advantage that AI offers, it is sometimes very frustrating to see if I know for internal tools how fast we could be and I would love to have that same speed in like our core product but I know if you do that we run the risk of disappointing our customers and I think this is not worth the speed advantage.

 

Hannah: Definitely, definitely. Well I think the the product is is very special, it's very amazing. I've loved seeing how it's become successful in more and more different countries and markets. if you are a real estate agent at a portal, best way to get in touch with you, Florian, is how?

 

Florian: either through our website, geosci.de or write

 

Hannah: GEOSCI.DE I'll put the details on there.

 

Florian: Yeah, LinkedIn. I'm very active on LinkedIn. I try to share relevant content on AI property search data or anything that's Hopefully find helpful. So that is another way.

 

Hannah: Great. Well well hopefully lots of portals will be bringing all of your data in and creating this amazing experience for all of their their customers. thank you so much for joining us. It's been super insightful, very honest and open, which we we love we love those those kind of answers and questions. So thank you very much for your time. Matt, any other questions for

 

Florian: Thank you for having me.

 

Hannah: you?

 

Matt Gayleard: No, I I think I think it's been a great conversation and and I think getting to be part of this podcast series is is incredibly selfish. we just get to meet interesting people and learn lots of new things. So it's it's been great. Thank you.

 

Hannah: Well, Flora, you are definitely a

 

Florian: I laughed as well. Thank you.

 

Hannah: interesting person to to meet. So thank you once again.

 

Florian: Thank you.

 

Hannah: and we look forward to to seeing your success and seeing lots of your data very soon. Thank you.

 

Florian: Thanks a lot.

 

Han end pod: Thanks for listening and thanks to Matt and all the team at iovox for the questions. If you're working in property or tech and want help with growth or product innovation, then talk to me, hannah at inoki.agency Otherwise, see you next time. subscribe and like.

Sponsor
iovox
Florian Gruenwald Profile Photo
Florian Gruenwald

Founder / Data Expert

Florian is the founder of GeoSci, a location data business based in Hamburg. He started his career in management consulting, moved into tech as a Scrum Master, Agile Coach and Product Manager, then founded GeoSci as a solo founder. On the show, Florian talks about what Europe’s location data can teach other markets, and what it takes to build a trusted data business today.

The Property Stack

The Property Stack is a podcast for the people building the future of property, from the world’s top portals to the estate agents doing deals.

Each episode, host Hannah Parker and co-host Matt Gayleard sit down with interesting people across property, tech and product, in the UK and beyond, to talk about what’s really changing on the ground. Less AI press release, more of what’s actually making a difference today.

Hannah brings more than 15 years of property and proptech experience to the conversation, including time at Zoopla, Hometrack and Propertyfinder, and is the founder of product consultancy inoki. Matt is Chief Commercial Officer at iovox, the data and AI specialists who sponsor the show.

New episodes drop regularly. If you’ve got a story or insight about what’s changing in property and tech, get in touch. We’re always looking for our next guest!

¨¨¨
You can find us on YouTube, Spotify and Apple Podcasts, or follow along on LinkedIn, Instagram, TikTok and Substack.

Got a story worth telling, or know someone who has? Give us a shout at thepropertystack@inoki.agency.

For all media enquiries, contact kate@inoki.agency.

  • Episodes
  • Leave a Review
  • Contact
  • Sponsors
  • About
  • © 2026 The Property Stack
  • Podcast Website by Podpage