SPEAKER_01 0:00

Commercial real estate is such a slow adopter of technology. If we think about the OCR technology, right? Optical character recognition. It's been around since the mid-70s. And there are companies that will have a picture of their least agreements, but they've not run it through an OCR engine to actually digitize it. So they're such a long road before they can apply any sort of AI stuff. Commercial real estate companies who will start thinking of themselves as data businesses will be set.

SPEAKER_02 0:32

Hi Karen, welcome to Built to Form Podcast. Thank you for coming on today.

SPEAKER_01 0:37

Hello and thanks for having me.

SPEAKER_02 0:38

What I'd like to do is really just dive straight in there with who are you and why do you do what you do?

SPEAKER_01 0:44

So I'm Karen. I'm the CEO and co-founder of a company called FIMA. We use existing security cameras that are already in buildings. We tap into the video feed that comes out of those cameras and analyze it with AI, telling you how your building is operating, how many people are here. We've just launched a really cool feature set that allows you to drill down to like a chair level. And in gyms, for example, that allows you to figure out whether your treadmills are being used, whether your free weights area needs upgrading or downgrading, depending on how much it's used. We have the first week's sort of data in actually from a gym. It's from a co-working space in London. Their free weights area is utilized every day, nearly every hour, even. And the two rowing machines haven't been touched once in seven days. So very clear way of okay, let's take them out and expand this for a bit. So that sort of stuff is what you can do with the data that we provide. Our AI has never seen a human face. And why we do it is FIMA started actually out as a consultancy, and we did all sorts of AI applications. I've been in the field for 10 years, which is, you know, way before anybody knew what an LLM was, or before they even existed. And we wanted to get away from people installing a million dumb sensors that don't get smarter over time and help them use the technology they already have and make it smarter and just quicker to use and install. So it all came from that kind of idea of let's make AI just simpler, easier to use, easy and quick to deploy. And we try to get away from all the jargon as well. Because again, if you rewind five years, maybe six years even, AI was expensive, like all the big companies were doing it, Apple did it, IBM did it, Microsoft did it, but it wasn't really just something that you could buy off the shelf. It's really changed, and we wanted to be there in the beginning and just really demystify uh AI as such. So we've always tried to sell analytics and usage analytics over AI and computer vision.

SPEAKER_02 2:51

Really interesting. And so obviously, demystifying AI and bringing AI into the market, is that why you're doing what you're doing, or how did you get into that? Because obviously, looking at your background, and we spoke off-camera about your background in terms of it's been quite wide-ranging. Why commercial real estate, and in particular, why commercial real estate and CCTV AI?

SPEAKER_01 3:13

Um, I've been in tech uh now since 2006 when I started with British Telecoms. So quite a while, dare I say, 19 years, oh my god. Um and I did a year and a half out of the tech world and worked in film and television, but I've I've been in tech kind of since. Uh in the very beginning, I was more ops, finance, that sort of stuff. And then from 2014 onward, to the past 11, 10, 11 years, I've been doing the business development basically and helping different kinds of innovative products go to market. Um why this specific field? Um, so computer vision. So when we did the consulting business, we tried to work with anybody that had enough data to do AI on, because that's something that you really, really needed. Uh, and with the application that we had developed for a city where we used existing security cameras just to figure out traffic flows and what the kind of potential carbon footprint is of a junction, we realized that there's what, 600 plus million cameras worldwide at that time. It's probably doubled by now since COVID. And nobody really uses that kind of data unless something happened, and then you rewind the footage and see kind of what happened. So we thought that why not take this into real estate in an operational capacity, not in a security or law enforcement capacity, and really help the world along post-COVID because we knew that patterns are gonna change, are we all gonna go back? Human-centric design needs data to operate in. Um, and I think a lot of people are very emotional about their environment, right? Whether it's the office environment, home environment, or the city they live in. Um, I am too. So putting that sort of data to a good use in terms of redesign um and operational management of the spaces that we inhabit is just something that seemed a very logical thing to do. So that's kind of why I'm in there and not doing very techie or defense or or any of the other stuff that we could be doing with computer vision.

SPEAKER_02 5:19

And so FIMA is an evolution then uh from consultancy to now what it is today, which is its own standalone AI and reporting solution, I guess. Is that the best way to describe it?

SPEAKER_01 5:31

Yeah. So the consulting business we we completely shut down. We did everything from time series prediction, sentiment analysis, uh, got uh predictive maintenance on a tenor port, like all sorts of stuff. Uh, tried to detect fraudulent applications for iPhone, like dropped a lot of people would drop their iPhone right before a new model would come out. Then you're trying to figure out if it was legit or or not. We worked with a couple of media companies, just anywhere where you could have enough data. Um, and then one of those ideas and projects that we had, we then spun into a completely new business, which is now FIMA.

SPEAKER_02 6:07

Specifically, the problem it's solving for commercial real estate landlords. What is that problem?

SPEAKER_01 6:13

Essentially, what FIMA is solving for is how many people are in my building, what's happening in my building. So, very often, I was quite surprised in the beginning. Uh, lots of landlords are blind to how many people are in the building, or they'll know how many people came in, but not when they left. Tenants don't always share what happens in the tenant realm and how many people they have showing up. So you're trying to figure out how much do I staff for? Am I really busy on Mondays versus Wednesdays? How big's that delta? If somebody's not consistently having tenants turn up, do I have to worry that they're going to churn as a tenant? Do I have to already plan for that? So it's just desire to figure out what's happening in my building and the lack of data within buildings. Um, the second uh problem that we're solving is just access to that data. So very often you'll have to install sensor technology. Uh, they don't sensors don't work if you've got a bit of a weird space. If the ceilings are really high or doorways are very wide, or you have both indoor and outdoor spaces kind of intermingled. Um, computer vision is very logical and works really well in those kinds of complex environments, and the camera's already there. So we kind of take away the pain of having to procure an entire hardware solution, test it for accuracy, and then install it as well. Because that whole process could take a year. We can go live in 20 minutes anywhere remotely in the world. So speed, quality of data, um, and the data itself. So often with AI, you need data to even you know figure something out of that data. So we help um uh landlords generate data out of their buildings that they can then use to query to figure out what's happening in the space.

SPEAKER_02 7:57

Yeah. So I think that's quite an interesting point as well, because the data's actually already there, it's just not being utilized in in a way which is usable, right?

SPEAKER_01 8:06

Yeah. So data's not all always there. So my point is that we rather than taking disparate data sets that commercial real estate companies may have that that are often not even digitalized, right? Uh we generate data for them, proprietary data that all belongs to them, that they can then take as a data set. And if they want to create their own LLM, dump it in there and ask, okay, I've got an office building in Clarkenwell, another one in Toronto. What's been happening in here in the last two months?

SPEAKER_02 8:38

Yeah, and what I mean by that though is the data is there in the form of a camera feed, right? It's it's already sat there. It's just not being utilized, it's not creating a data layer for value. It's just about creating value. I think that's what's really quite exciting about it is it's utilizing what's existing in a cost-effective way to then capture and create data which is then usable and actionable, right?

SPEAKER_00 9:03

Yeah, absolutely.

SPEAKER_02 9:04

Um, on that and data, I think it's worthwhile talking about because it's an emotive subject, right? Is data and CCTV, AI, those those three things put together can scare people, right?

SPEAKER_01 9:18

Yeah, of course.

SPEAKER_02 9:19

Um can you just explain how you go about resolving those those fears which come with that in you know, with GDPR and all those other different concerns which a landlord will have and tenants will have around how that data is captured and used?

SPEAKER_01 9:35

Yeah, of course. So uh we started the company, well, obviously in a European space that's governed by GDPR and the equivalent uh legal framework within the UK. Um so we set out from the very beginning not to do facial detection, gate detection, any personally identify identifiable information detection. Uh we don't save any video footage. So if you imagine a FIMA being like a sieve, the video feed passes through it, it's not saved, and only the metadata is kept. There's nothing to rewind, um, there's no phases to really look at, and uh the detection happen, detections happen really, really quickly, and then the video data is deleted. Uh, we also the way detections work with an AI model that we have is it looks at your hand, torso, leg, it doesn't look at anything above a neck. So that the model's not been trained in that way. It never kind of captures um a human face or anything above the neck. So that means that even if somebody is partially passing through the frame, we'll know that it's a human being, we'll count them. So, similarly with vehicles, if we just see a wheel or uh or a windshield, we'll know that it's uh it's a vehicle. Um, so that kind of safeguards against it. All the data belongs to the clients, not to us, so we don't kind of license it out to them, like it's the client's data. Um, and we have placed our servers in Frankfurt, in Germany. So we work with Google and AWS both uh because Germany has been one of the strictest adopters of GDPR, so we decided to put our money where our mouth is and stick all of our processing in there.

SPEAKER_02 11:17

Yeah, I think um an interesting viewpoint on that as well, and with with regards to how you're capturing that data, I think people do get quite emotive about it, and being able to explain that in a in a very logical way to them, exactly how that's going to be captured, how it's gonna be looked at and not retained is is really important. Um in terms of the planned space, obviously you said it's a it was an evolution. So when did you decide that actually commercial real estate was a route forward for FIMA?

SPEAKER_01 11:56

Just go given the previous um question as well. So sometimes in a deployment we don't go to tenant spaces and we don't go to meeting rooms and really small spaces like that, again, because there's a comfort level that the people will have with camera technology. And in that case, a hybrid solution where you have a sensor technology deployed in a space like that, and FIMA in public areas, mixed use areas, more complex spaces, probably make most sense. Um, so how FIMA got involved in in commercial real estate space, or how we kind of started out. So the very first application, like I mentioned, was for a city and to try and figure out what's going on in a junction. They also were using sensors. Um, one sensor could only figure out what's happening on one lane of traffic. Uh we didn't look at pavements, uh, and the sensors were able to that place still uses them, but that city still uses them. They would look at uh short vehicle, long vehicle, and could only really determine car, bus, and truck, maybe a motorcycle. So we do car, bus, truck, motorcycle, e-scooter, tractor, human. Uh we have um separate micromobility models, so we can look at how many cargo bikes are in a city, for example. Um, if a client needs a specific deployment, we'll create it for them in the space of like a day or two. So we can look at a lot richer data set. We launched the company into COVID. So we properly launched like early 2020, yeah, and everywhere where we wanted to go live, which was public domain, was completely devoid basically of people. So it was pretty difficult to do any sort of light testing. So we spent two years and a bit more even on proper RD, got a patent sorted, brought the own cost down a lot. So when we started out, just paying for GPU resource was 250 euros per camera feed. That was not something you could sell to a customer and have a markup on it, right? So we were able to bring it down significantly. So a lot of that kind of work was going on. Um you also realized at some point that selling into public um spaces where we would be very useful for a local government authority. Like let's say I live in Cambridge, and they've put in a scheme near where I live where a bridge has recently been closed to kind of car traffic. So only taxis and buses can go over it. Huge debate in the local community about whether that's good for business, bad for business, is it more pedestrians now or not? You could just deploy us for a month, have a look, it's really cheap, and you'll know. And then you can go to your next election cycle with data. Um, but going back to data and public uh spaces and environments, it's very emotional and very very politicized. And we realize very quickly that it's just our runway is not long enough to have debates with uh local authority politicians. So we completely pulled out of any sort of analytics that happen in roads or kind of public areas and switched into um commercial real estate because that's where the demand started coming from, basically.

SPEAKER_02 15:02

And you're talking about AI and FIMA. What for people listening and watching, there's obviously different levels of AI, and it'd be great for you to explain that to us in a layperson's term, because I know you've got the ability much better than I do to do that, um, because I'm the layperson. And then also um what FIMA actually uses as well and how how that works. Not asking for IP though.

SPEAKER_01 15:26

Yeah, yeah, no, no, no. Um I think when we talk about AI now, most people mean a chat GPT or Claude or some other like large language model. So they've kind of blended into the same term. AI is a much broader term, as such. I actually really recommend if somebody's interested in a book uh called Um it's by Michael Wildridge, he's a professor at Oxford University in in AI and maths, and he has written a book about the history of artificial intelligence, and it's fantastic just to get to grips to where the term came from and you know how long have we been in the field. Uh a surprisingly long time, isn't it? Yeah, I mean computer vision as such was first launched as a field of research in 1954, which is a long time ago now. And scientists were trying to figure out how cats see, like is it blended edges, sharp edges, and how do we then figure out how to make a computer also see? So that's you know, it's evolved in different ways, but thanks to LLMs, it's just accelerated, and and then people kind of you know now query what is AI and what does it mean. Um there are AI systems, obviously, there's machine learning that's been around for a long, long time. Uh, there's agentic AI that we now talk about. There's also automation. So I often find that people confuse what is automation with AI. So that gets blended in. Automation is when you uh manually create the sequence and then basically let that run automated. For truly agentic AI, there's a company, for example, out of Estonia, um they have agents that work alongside your procurement team that have their own budget, right? So that's an agent that can go and negotiate contracts within a limited scope. And that's truly agentic AI, which we I don't think are seeing in in the world really yet, uh, properly. So it kind of started from that narrow field and then it's now kind of moving up. It's it's a funny one. So a lot of uh AI people or practitioners don't use the word AI. We would use big data, machine learning, computer vision. Like you talk specifically about the type of AI you're building or working with rather than using just the term AI.

SPEAKER_02 17:45

Yeah. And say you're utilizing computer vision.

SPEAKER_01 17:49

We're utilizing computer vision. We're also starting to use a little bit of um, well, starting to use. When first ChatGPT launched, we had FIMA GPT. So you could go and query anything about your data, but people who didn't know what to ask. It was really interesting in the beginning when that was just an empty box, and it was like, what do I ask? I think now it would be a lot easier. Uh, we're trying to get LLMs to um summarize data in a reporting. So basically you log in and you have an executive summary instead of having to look at five pie charts and line graphs and so forth. Uh, we're also looking at VLMs, so that's visual language models, which can interpret both text and image and the correlation between both of them. So currently LLMs are really good at text, but not so good at doing both images and text and reasoning within that space. So that's something that we're we're starting to look at specifically for commercial real estate.

SPEAKER_02 18:46

Yeah, and on commercial real estate, um where would you say AI is and where it's headed as well? Because typically we're a little bit slower.

SPEAKER_01 18:58

I don't think AI is really anywhere beyond a chat GPT claw type setup, maybe for marketing. Um commercial real estate is such a slow adopter of technology. So if we think about the OCR technology, right? Optical character recognition, it's been around since the mid-70s, 50 years. And there are companies that will have a picture of their lease agreements, but they've not run it through an OCR engine to actually digitize the content of that lease agreement to then be able to put that into a system and query it. So there's such a long road in a lot of these companies to truly digitize themselves before they can apply any sort of AI stuff over it. Um and and you know, rubbish goes in, rubbish comes out. That that's kind of the way. So where I've seen commercial real estate do well is where there's one large UK company that has given data guardianship roles to various company, various basically individuals, all the way from CEO down. They're either data guardians, data um uh I can't remember all the terms, but everybody has like a specific role. Either they look at governance, they look at how to better share it, they look at uh privacy, they look at ethics, and then because they're cross-functional, they're able to now bring a lot of that different knowledge together into working groups and figure out how to make more of their data. And I think those kinds of companies will be successful. And there's another one of our clients in Canada that's working at creating a proper pan company data lake with every single possible data source going into it to then start querying it with AI and trying to figure out are there any anomalies anywhere? Everything from leases to occupancy to uh staffing, you name it, like what is happening in my organization. But they've been on a I think they've been doing that for about two years, just trying to get all of that data into a format where it's even queryable. And they're a massive company, they have a they have over 70 billion dollars worth of assets under management, so they have the resource to do that kind of work. Um, companies, commercial real estate companies who will start thinking of themselves as data businesses will be successful. Nike has said for years that we're a data company. They don't they don't say that we're like a retail business or just a shoemaker, they're in the business of data, data about how you buy, how you perform, how you run, and then they build a product to match that. So I think that shift in thinking it really still needs to happen.

SPEAKER_02 21:37

Yeah, I'm seeing green sheets of that, but going back onto the AI part in commercial real estate and probably outside of the um etc. as well, and kind of what you said earlier is people in the AI field, they'll talk about their specific AIs, but there's a bit of cloak cloak and dagger, not cloaking dagger, but um grayness. To some people claiming to be AI when the reality is their products aren't actually utilizing AI as well, which is something we've got to be quite wary of as well at this point in time.

SPEAKER_01 22:11

Yeah, so this is going to the time before we launched FireMark proper. But we used to do diligent due diligence for a uh for a European venture capital fund. And out of eight companies, six did not have any uh IP that would kind of hold up. Uh you know, we have proprietary AI, they didn't have any. Um and yeah, you're right, a lot of companies probably don't have it. That being said, if somebody repackages an LLM and kind of tweaks it and trains it, or uses it, or builds something with, I don't know, lovable and people use it and it brings value, that's I think it's fair, like it's a fair use case. So their IP would be rather than creating something, I don't know. So I've got a law diploma. So rather than creating something completely unique and patentable, they've created a useful model, for example. So they've taken some of that that exists and put it into either a better use or a novel uh use, for example. But yeah, there are there's a lot of like for I mean we both go to a lot of these prop tech conferences, right? And you look at the panel and it's full of AI people or or topic is AI, and you know that everybody has bought that seat uh for themselves to be on that panel, and they're not industry experts. And I've heard all sorts of absolute just rubbish being spoken from stage that's just not factually true. And then you're like, I mean, come on, just some conference organizers need to really do a better job at getting like actual practitioners on stage rather than just PR people who are trying to sell themselves an AI company when they're not.

SPEAKER_02 23:54

Yeah, I think and that's doing AI in the industry a disservice and gonna has caused, I think, a level of distrust, right? Yeah, um and you know your company might be utilizing it automations, just be honest about that. That's if it's still doing a good thing and it's adding value and it's bringing something, you you don't need to jump on the AI bandwagon and say we're utilizing AI. Be honest about it. If you're using machine learning, talk about machine learning. Let's actually inform people about the technologies which are being implemented rather than trying to put it all underneath this AI banner because it's an exciting and shiny object at the moment, right?

SPEAKER_01 24:32

Yeah, absolutely agree. So there's a mom, uh, one of some I've got two daughters, and the younger one, uh, one of the mums um from that school works for a pharma company in Cambridge. And she's a data scientist, and she did kind of sigh at me and say, Oh, you know, all the management people want AI this, AI that, then we can just use a good by easient model, and then you know, it's quicker, easier, and cheaper to run. So it's not a panacea for everything. I'm also a uh a panelist uh regularly at the Cambridge University Judge Business School Gen AI course. So they have a course for generative AI for executives. Um, I've been invited in since the beginning of the year, and when I kind of a lot of that kind of uh time that we spend together is also discussing things. And often people say, and this is from just a few weeks ago when I was last there, top management again is putting pressure on the organization to adopt AI and agents and do stuff. Uh, often they don't tool people properly for the task, so that's a major issue. They don't augment their workforce with the people who come from the industry or have the knowledge or have it implemented elsewhere. Um and I think they need to do that. Upskilling is a huge just gap that's just not happening, especially I think in real estate, where I don't know, facilities managers are maybe trying to do stuff with AI and then they just need the tools. Um, and and that was also, you know, the comments are, but there are better better systems already in place. Like, why do we need AI specifically for this? Is it just so that you can go and and say that we use AI or like exactly? So if I rewind again six years, and we were a consulting business, um we would always ask a client, have you tried every other way? Have you tried everything else? And if you've not, go away, try all the other ways, and then come back, and then we'll see if like AI can help you and we can build a model for you and and and see if you've even got enough data.

SPEAKER_02 26:31

Yeah. And I think that probably comes down from it's a shiny big object, but then the end clients are saying, Hey, what are you guys utilizing in their in their RFPs, for example? What AI are you using? And then it's this, oh god, we need to have AI just to pass this RFP process. And so yes, people are scrambling around saying, How can we do that? Whereas actually the majority of people that are still at that point in the industry, and I think we've I've talked about this on previous podcasts as well, is most people in commercial real estate are let's get the data clean and in a usable place first position. And I think everyone's got to not probably stop worrying so much about what other people are doing and just worry about getting their own house, their own house data in order first, and then you can work out what processes or systems you can apply before you're looking at the really fun AI stuff, right?

SPEAKER_01 27:22

Yeah, I mean it's in a way quite typical business transformation program that you're looking at. You go in, you figure out what problems you're trying to solve, and then you see whether an AI tool is a match for that or not. And it may always not be a match for it. Um, I mean, we're not gonna put AI back in a box, it's it's here to stay. Um so if I again if I look back at the Cambridge Uni stuff, everybody gets a pre-course uh survey. And in the beginning of the year, when you looked at kind of adoption of AI tools, uh companies and people were at the kind of zero or one. Like some people kind of tinker about with it, some organizations, individuals use it, but the business is not yet using it, etc. Now, the program survey I looked at two weeks ago was uh companies are already using it or already have developed a proper strategy. Uh very few, like very few, maybe one or two, were where you know, at a the stage where the people feel like we have a proper mature strategy and it's being implemented and we're skilled and so forth. But it's kind of moving that way. And a lot of people are daily uh and weekly logging into either ChlorChat GPT, another LLM. So it's being used. Um but yeah, it's kind of one of those like it still needs a bit of time and guidance uh to be properly properly adopted.

SPEAKER_02 28:44

Yeah, and on that, how can real estate leaders you know try and get ahead of the curve and start you know implementing it?

SPEAKER_01 28:52

So the companies, whether real estate or not, if the CEO is a champion, it usually gets adopted and kind of prioritized and you know it goes better, better or well in that kind of business. Uh where it's one of those top-down things that deal with it and the CEO just wants to see results without using it. Um it doesn't go that well because it's it's about creating trust. A lot of people worry about their job, frankly. Yeah. Uh whether I will be replaced or not, uh, whether they're gonna reduce the workforce because we can do like 1.5x more work thanks to AI, etc. So one really huge thing is to create a culture of safety and trust that this is a normal thing to use. Um, a friend of mine gave an example, one of the companies he works with, the CEO has a Chat GPT box open on his kind of screen all the time, and people know that he uses it daily. So they don't worry about using it. Because otherwise, what they'll do is they'll take the data, use it on their private account, and all of your proprietary stuff leaks into the internet and gets absorbed into the models. So that's really relevant, is to be kind of clear around it. Um, and if you don't know what you're doing, get help. There are consultancies out there, there are companies who've done it. Um, just get somebody to come and help you with implementing that. And a huge piece is education. And I think that's something I'm gonna keep saying for for a long time. Educate yourself, educate your people, tool them, uh, augment your workforce with people from the industry, and then you can start adopting it.

SPEAKER_02 30:23

And that brings us me on to a point, and you you've touched on it twice now, is something where you've you've done this and other other people I've spoken to on this podcast as well, being from outside the industry. Um, and it's this acceptance, it's bringing people into an industry which is typically being quite closed off, which doesn't help with us changing or moving fast, um, but has been closed off to outside influence. And I think now's the time, right? That you're only gonna be able to we're only gonna be able to learn from outside the industry, especially on this subject.

SPEAKER_01 30:57

Funny one. So I have a we have a really cool client. Actually, the first company who started using us is a mixed-use uh campus in Tallinn, and they needed a solution for their parking because uh they have nearly 2,000 spaces spread across five car parks. And Estonia is a place where you have sunshine all of June, maybe sun, it's dark only four hours a day. And the winter it's the opposite, it's mostly pitch black, except for 10 a.m. and you know 3 p.m. Um, snow, so you can't put sensors in because snow covers them. Constant issues. So we went in and kind of helped them figure that out. And they had brought in a CEO for a big, big commercial real estate kind of company from the tech world and a woman at that. So she was initially ahead of innovation and then became the CEO. During her tenure, that campus grew to encompass not just offices, they have a university now, they have um a daycare, an international school, a building that's dedicated for health and wellness with private clinics in there. Um the tax and customs board relocated their headquarters there. Uh, they're moving away from cars completely. So they want to be completely car-free as a district in the next, I can't remember, five or ten years. Proper ESG strategy with the data visualized to every tenant every day on screens across the office. And now, I mean, we've been working with them for for five years because they literally took all of our MVPs and put it to place. If they were a city, they would be the third largest taxpayer in the country. So there's Tallinn, there's the next town, and then it would be them. So they've been very successful because somebody with a completely different mindset's come in, looked at it, and thought, no, no, we can innovate in a very different way. Um, and they've put tech in from the very, very beginning.

SPEAKER_02 32:51

That's really cool.

SPEAKER_01 32:52

They're a really cool business.

SPEAKER_02 32:54

And I think it brings me very nicely onto a little segue which I like to talk about. Um, and obviously you mentioned female CEO, um, and I know it's something which you're obviously really passionate about, is getting more women into STEM, but also into managing roles, right? Yeah. Um and so on that, how have you found commercial real estate? Do I ask? Um as an industry for that.

SPEAKER_01 33:21

Oh, so I don't come from a STEM background. Yeah. So that's one thing I'm trying to tell women is that it's not a barrier. It could be depending on where you go, but it's not necessarily a barrier. Um, my job is about uh relationship building and people. And I think women are innately very good at relationship building and with people as well. So that's kind of uh where I try and come from. Commercial real estate industry, like the VC industry and all sorts, is very male-oriented, very male-dominated. Uh, I'm just off uh you know, a US trip where we went to a bunch of conferences, and it's very, I mean, our computer vision model would get very confused there because everybody is in a dark navy suit, uh, white, and of a specific age and height. Um, so it's very homogenous. The industry is very homogenous. I know that there are spaces in which I wouldn't be welcome, but our head of sales is welcome into and will do really well there. And that's why I hired a man to do sales for us because you know that you know I can close deals, but I can't get into all of the spaces where I want to be in. So you just kind of make allowances for that. Uh, and you can't be angry about it. It's just the way the industry is. It maybe we'll change, maybe not, I don't know. Um, and there are women groups kind of popping up, but globally women's rights are being rolled back, and I think it will reflect in in the workplace as well. That it's pretty tough. Um, as an industry, I find in the UK it's a lot more um, I think, open and accepting, and just it's a very tolerant society anyway, compared to maybe some other places. I'm originally from Estonia, a lot more parochial, tiny, you know, Soviet Union dominated for for a long time. Um, so that's a little bit maybe different. So I feel better here. Yeah, but still obviously there's lots of restrictions.

SPEAKER_02 35:18

Yeah. So yeah, I think we're probably talking more about commercial real estate in the UK is potentially a little bit better than say America or better than America, maybe in some of the It is better, yeah, yeah, yeah. And some other parts of the the world. So that's positive. But I think how can we improve and what needs to change and why?

SPEAKER_01 35:40

For tech companies specifically, uh, more women need money. So we don't need mentorship, we don't need sponsorship, we don't need events where people teach us stuff. Like we need hard cash. Like women need money, just like uh men are funded for the most sometimes ridiculous ideas. Uh by the way, a lot of FemTech, so uh products that are oriented for women or period products or whatever, are are funded or kind of operated by men. You'd think that women would be able to raise money for the for you know that kind of stuff, but uh it's tough. So I actually have a t-shirt that says I'm the effing 1%, uh, because only 1% of women uh raise money, venture capital. So if you're a solo female founder, it's very unlikely you'll raise money. It's very, very tough. If you're a mixed gender team, it's gone from about 10% to about 16% capital.

SPEAKER_02 36:33

Yeah, latest stats I saw as 14%.

SPEAKER_01 36:40

And the rest of the money goes to solo male founders. And when you look at Atomico or any of the reports that come out, it all laments the same stuff, the same statistics get churned out year in year. And then you look at the portfolios, and I I have I have manually Googled through a bunch of these portfolios for European VC firms, and they don't put their money where their mouth is, and they absolutely overwhelmingly uh uh finance um male, only male teams and kind of male founders, and it's just really sad. But that means more women need to be making the decisions. There's not enough partners at venture capital firms, and if you're an LP limited partner and you put money into a venture fund, you need to hold that venture fund accountable for making sure that uh the funnel they're getting is more equalized so that the investments go both to men and women. Because we know that women bring back more money per dollar invested than men, and that could be just because the barrier for us is so high that only the real, real top ones get across it, and then of course you're better. Um, also, women do usually more funding cycles because we lay raise less money. We need to raise more, you need to go through more cycles, which means you dilute yourself more, which means your exit is not going to be as great. So there's so many systemic kind of trips in there. It needs to be, I think, addressed from a regulatory and culture point, but also VCs need to look at the funnel that they're bringing in and have just more female partners. I think that that's really, really irrelevant for tech businesses.

SPEAKER_02 38:17

And is that is that what leads to the bias, or is there is there other things which you can think of which are causing that bias towards an 85% bias towards a male?

SPEAKER_01 38:27

And so some of the research that kind of comes out shows that when you show people identical uh portfolios or identical companies, one's led by a man, one's led by a woman, a a male committee would prefer the male. And it's kind of because that's what we know, that's we that's the spaces that we move in, that's kind of where we grow up. Even in comparison real estate, like golf days are a big thing, right? A bit every year there's a group that does a cigar and a whiskey thing, which is like you know, obviously quite a male thing to do.

SPEAKER_02 39:00

I hate both Justin Viva.

SPEAKER_01 39:01

Yeah, I don't smoke and I don't drink. Um so it's one of those ones where if those spaces scream male, then why would I why would I go and attend and and then I don't even cultivate a network where I should be? Or if I do show up, I'm usually you know one of two people who turned up uh who are who are who are women. So I think that's spaces need to change a little bit for that.

SPEAKER_00 39:26

Yeah.

SPEAKER_01 39:27

It's ultimately we will fund what we know. Our networks are very often gendered. You know, men will hang out more with men, women will often hang out with more with women. It sounds really odd to kind of say this out loud, but yeah, there's so much research around that.

SPEAKER_02 39:42

Yeah. And I I know you said about the mentorship and all that. That's like don't need that, need money. What about allyship?

SPEAKER_01 39:48

Allyship is great. Absolutely, that's great. But allyship also would mean you going, you know, pushing, pushing from women to get the financing, to get the contracts, to get the introductions, uh, and to be in the spaces where those deal makers are and the check signers are. Like that's allyship.

unknown 40:08

Yeah.

SPEAKER_01 40:08

Not just saying, uh, you know, let's pause with Karen and do a LinkedIn post and do a hashtag allyship. That's not allyship.

SPEAKER_02 40:15

No. It's a really tricky position, right? Because I I'm kind of everything we're talking about, right? And also it's something which it is something as well, because in the sense that we I talk about this with my wife and whatnot. We have we also have this problem as well that that men are also more likely to put their hands up for something and say that they can do something when they can't, as well, and they are they're more willing to kind of not willing, they're they're ready to fluff it. They uh whereas women will be more measured about kind of how much experience they have and take that into consideration more than some most men are like, yeah, I can do it, if they've got like 10% of the knowledge and they think they come back for the 90%, whereas women are like, I need to have 80% of the knowledge in order to get that role, if that makes sense.

SPEAKER_01 41:11

Yeah, I think some of that is cultural as well. So um if I look at so again, when you look at kind of research and and you take a group of men and a group of women, then there's bigger differences between the extremes of the men and women within sorry, within the men within that group than between the two groups. So if you just look take men and then one end of the spectrum and the other one, that's a really big gap. And and to the average man and the average woman, like the gap is actually not that huge. Um, also culturally, if you look into the US and some of the crazy high-profile criminal cases we've recently had with female founders who have taken risks, like the bad blood one, for example, Theranos. There's another woman that was just recently sentenced to jail for fabricating uh clients that they never had. Um so there are women who who do that kind of stuff. Uh, we usually get punished more for that. So if you look at even Adam Neumann, who I mean, I read the book and the articles and seen the dramatization of it. I I'm shocked that he still gets money and gets funded after after we work um and cashed out in the tune of hundreds of millions, you know, disappears for two years and comes back on the scene. So so women, I don't think, would ever be allowed that kind of leeway. Um uh and also we're again culturally depends, right? I was raised in a in a farm. So again, culturally, Eastern Europe's different from the UK and then from the US and maybe from Spain. So I think a lot of the gender roles we carry and embody obviously come from childhood and the culture that we're raised with. Um, so if you're always so girls are um commended for being quiet and compliant and having neat handwriting, uh and then you know the whole boys will be boys thing. I think we're more aware of it, yeah. Uh like definitely more aware of it, which hopefully will help. Um, but yeah, the way I was raised and the way I'm trying to raise my kids is very, very different. Yeah, very different. Because I mean I was the one that had to like uh mop the floors and do dusting and that kind of stuff. And my brother was never subjected to any sort of domestic cleaning tasks. And I found that really unfair. Uh it is it is unfair, but it was just the way that boys just didn't do that. So I think culturally that's a more of a Eastern European thing uh than would would have been the case here.

SPEAKER_02 43:42

So moving on, probably back into the AI world, we obviously talk about AI and well, commercial real estate being slow adopters. W why do you really think we have lagged behind the other sectors so much? What is it about commercial real estate which Causes that lag?

SPEAKER_01 44:01

I think commercial real estate specifically has been very um. I mean, it it's just a slow process to build a building. Like just to put a building up. It could take decades to get through the planning permission and however many different design iterations you will have done by then point and then build it, and you could run out of money halfway through that. Um there's so much risk in every single step. Uh, returns can be completely eaten up by the end of that cycle. So it's very natural to be so conservative when you're investing and putting to use tens of millions of investor investor money, and it has to play out. It's not like in a startup where you're just taking a bet, right? Um, so I think that's one of those things where it's just when you're a guardian of money, uh, you don't run after every new tech innovation that's probably out there. Um so especially hard for a startup would be to sell into a business like that. Because if you've not been around for long enough, you're just not trusted. Like, will you still be here when the building's up and operational? So then I think it's a very fair question. Um, the other thing is probably just uh, you know, you need to sell square foot or square meter, and everything else trickles down from that. So, unless there's a very, very clear and immediate ROI, it's very difficult to adopt a tech solution. Like so what it's AI. I think that's a really, again, a really fair question. And more companies are now questioning the so what aspect. So again, with the Cambridge uh core stuff, in the beginning of the year it was more a cool gimmicky marketing text, cover letters, speeches. Now it's the okay, but so what? Like we have 50 licenses for whatever tool we use. What's the val what's the actual value of it? Like, where where can we really innovate with it?

SPEAKER_02 45:56

Yeah, I do find that quite interesting in part of commercial real estate as well, which obviously I'm involved in with iBosses. And from being in different parts of the sector myself and being involved in very different sectors of commercial real estate has surprised me just how focused prop tech is on ROI. As in people don't look at ROI in the same way at other parts of commercial real estate, like they insist on looking at prop tech as okay, what's your ROI? That's a question which doesn't get asked of really in other parts of commercial real estate to the same same extent, which I think is really quite fascinating. I think why why are we so obsessed with this this ROI on prop tech, but not obsessed about it in terms of maybe an FM solution or something else, you know?

SPEAKER_01 46:51

I think it's how I mean PropTech's new, like it's it's a pretty new field, right? So I think there's a little bit of distrust maybe there. Uh it's an easy question to ask. I I think it's just an easy question to ask to then be able to go and justify that somewhere. And now fundamentally the whole thing comes down to cost. Who's gonna pay for it? Is it coming out of the service charge? Am I gonna pass it on to somebody? Who am I gonna pass it on to? Uh I want to look good in front of the investors, the board, the stakeholders, the shareholders. And if we blow the operational budget out of whatever proportion, then I'm gonna look bad. So, you know, but if I can go and say every penny you put in brings 2.5 back, gonna be amazing. So I think it's just the reduction of uh debate and purchasing decisions to a couple of very, very basic checkpoints.

SPEAKER_02 47:44

So um closing up, um I've got two questions which I tend to like ask my guests. Uh the first one is big picture looking at commercial real estate, what needs to change and why.

SPEAKER_01 47:55

I think we covered it a little bit as well, which is I think reskilling people, uh upskilling people, and augmenting your workforce with people who come from a technology background. Because it will not so it will be touched by tech and it will be touched by AI, and it's better to have people in your organization helping it get ready to properly kind of adopt it. So I think commercial real estate will move on from simple LLM queries to help their marketing brochures and will really start adopting it. There probably will be a couple of um not a couple, but I think there might be mid-market developers, asset managers who can kind of tear ahead from the rest because they're more agile and it's not turning around such a big ship, and that will give them a proper competitive advantage over others. So I think that the landscape might change a little bit.

SPEAKER_02 48:49

Nice. And then looking at commercial real estate, do you think it needs to evolve or uh revolutionize in its approach?

SPEAKER_01 48:59

To AI, you mean or everything.

SPEAKER_02 49:02

And everything. Let's look at the whole of commercial real estate. Should we evolve? Are we in evolve or should we uh revolutionize?

SPEAKER_01 49:10

I think uh oh god. So in the UK, uh it's hard. I think it's hard to revolutionize in a country that's never had a revolution, so it needs to definitely evolve uh to catch up with to catch up with things. And I think there's big regional disparities in on offering. So when you go to Scotland, for example, and you come down to London, even buildings obviously look different, they're differently used. Um but to really attract talent, like legit good talent from a wide range of ages and industry experiences, it needs to seriously step up and and evolutionize, or whatever the word is, uh evolutionize, like revel revolutionize and evolve together. Um, but yeah, I'm just thinking back to Blueprint uh a couple of weeks ago in in in Vegas and the Venetian. And yeah, there are women there. There's always a women's dinner that kind of takes place, but it's boring to listen to the same script year in, year out that says, let's get more women involved and more tech involved, and you know, just do it. Yeah, just do it.

SPEAKER_02 50:21

So actually, Ben Cross, uh previous guest which I had on, he he called it lots of mini revolutions.

SPEAKER_01 50:29

It's a mini revolution.

SPEAKER_02 50:30

Lots of mini-revolutions to evolve. Um it kind of sounds like a similar sort of viewpoint, really.

SPEAKER_01 50:36

Yeah.

SPEAKER_02 50:37

Revolutionise little parts of the sectors to get a bigger solution across the whole.

SPEAKER_01 50:43

Yeah.

SPEAKER_02 50:44

So no, thank you so much for coming on today.

SPEAKER_01 50:46

No, thank you, James. It's great.

SPEAKER_02 50:48

Really good to have you on. Really uh, you know, amazing topics which we got on, and I think we your passion, your insight into AI and everything else around commercial real estate is is really clear to see. Um, it's it's been really refreshing to hear as well. And I think one of the big things about this podcast is one thing which I always hope is that there's going to be some good value, some takeaways for people listening in here. And I think people really need to listen to this um episode. It's gonna be a lot of value for people to take away to understand how we can improve as an industry and really move forward, both in terms of the social demographic side of things, um, but also in AI and technology as well. So I really appreciate all your insights and information on that.

SPEAKER_01 51:38

Thank you so much. No, thanks. Thanks for having me.

SPEAKER_02 51:40

Thank you.