Latest / Built to Perform / Built to Perform with James Hallworth – Ep.2: Fredrik Wisloff on AI Transforming Real Estate

Transcript

Built to Perform with James Hallworth – Ep.2: Fredrik Wisloff on AI Transforming Real Estate · Built to Perform

• 0:00
  SPEAKER_00: How do you enable the organization to start using AI on a bigger scale?
• 0:04
  SPEAKER_00: Either you are embracing it or I think you are in danger of being extinct.
• 0:09
  SPEAKER_00: Everybody should look at how they can utilize this technology to augment them to become even more teacher.
• 0:19
  SPEAKER_01: Hi Frederick, uh, thank you for joining us on the Built to Perform podcast.
• 0:23
  SPEAKER_01: It's great to have you on.
• 0:24
  SPEAKER_01: Um let's dive straight in with who are you and why do you do what you do?
• 0:29
  SPEAKER_00: Uh thank you, James.
• 0:30
  SPEAKER_00: It's very nice to be here.
• 0:32
  SPEAKER_00: Uh so my name is Frederick.
• 0:34
  SPEAKER_00: I'm the CEO and uh co-founder of uh Findable, and we're solving the document pane within real estate.
• 0:41
  SPEAKER_00: So everybody who owns or operate real estate are very familiar with the pane of uh documentation in multiple formats in multiple places, and we're finally solving that problem.
• 0:54
  SPEAKER_01: Why did you get into findable and into commercial real estate?
• 0:59
  SPEAKER_01: Because looking back at your your history is it hasn't always been in the commercial real estate world, and you kind of moved into this world.
• 1:06
  SPEAKER_01: So do you want to just give me a bit of a background into the journey and why you're doing what you're doing now?
• 1:11
  SPEAKER_00: Yeah, absolutely.
• 1:12
  SPEAKER_00: So I have uh a technical background myself, but uh before we started the Findable, I used to work as the CTO for the leading real estate data company in Norway, and that's also where I had the pleasure of uh hiring two of my now co-founders, Lars and Knut.
• 1:28
  SPEAKER_00: Because back in 2018, uh we were working on a project on how to extract data from around a million real estate transaction documents.
• 1:39
  SPEAKER_00: And uh even back then you had out-of-the-box solutions which promised to solve this problem.
• 1:44
  SPEAKER_00: But when we started testing it, we discovered that it didn't really work.
• 1:48
  SPEAKER_00: And uh back in 2018, that was uh one year after the transformers var introduced.
• 1:55
  SPEAKER_00: So we quickly moved into using LLMs to solve this problem, uh, which was of course way pre-GPT.
• 2:03
  SPEAKER_00: But I was blown away by the capabilities of these large language models back then.
• 2:09
  SPEAKER_01: And I think it'd be really useful just to if we do some very basic AI understanding terminology, because I think um there's a lot of different types of AI and different levels of AI.
• 2:24
  SPEAKER_01: And so obviously you mentioned large language models, but you also um probably mentioned machine learning and other bits and pieces in it.
• 2:31
  SPEAKER_01: So, do you want to just kind of explain the a few different levels as to where we are with AI and kind of where findable is in terms of the AI model which you're utilizing?
• 2:41
  SPEAKER_00: Yeah, absolutely.
• 2:42
  SPEAKER_00: It's a big topic, but I'll try.
• 2:45
  SPEAKER_00: So I think you know, fundamentally, machine learning has existed for a long time and the ability for machines to learn and adapt.
• 2:53
  SPEAKER_00: And I think that you know what's uh become popularized now through AI and GPT was, you know, first available back or the Attention is All You Need paper released in in 2017, which really opened up the transformer-based model, which is the foundation for the generative AI that we're using today.
• 3:14
  SPEAKER_00: When we started Findable back in 2020, we didn't really have uh open AI and GPT, so uh we based it on different models, still large language models trained on massive amounts of data.
• 3:28
  SPEAKER_00: And then we basically fine-tuned these models based on real estate documentation.
• 3:34
  SPEAKER_00: So we've actually carefully curated over hundreds of thousands of real estate documents to be able to train our own models.
• 3:42
  SPEAKER_00: Uh, and uh so I think for the previous five years, it's been a crazy ride within AI and the launch of uh GPT.
• 3:51
  SPEAKER_00: And I think that so we're certainly utilizing GPT in our platform as well.
• 3:56
  SPEAKER_00: Uh but I think the part of our success recipe is that we're actually using the whole tool chest from old school image processing to machine learning and also generative AI in our processing.
• 4:10
  SPEAKER_01: In in that journey for you, obviously AI has grown, changed leaps and bounds in that time, like you said.
• 4:18
  SPEAKER_01: And in terms of your own uptake and people taking up Findable, have you found there's maybe more trust or that's opened up the market a bit more for you since everyone got quite excited on the hype around generative AI?
• 4:35
  SPEAKER_00: Absolutely.
• 4:35
  SPEAKER_00: I think it's been a massive shift.
• 4:37
  SPEAKER_00: I think uh you know when we started in 2020, 21 and approached real estate companies about using AI on their documentation they were skeptical, but we did have some early adopters who decided to take a chance and we were able to deliver amazing results for them.
• 4:56
  SPEAKER_00: I think during the real GPT boom, everybody believed that that was, you know, that was everything AI was, which basically is wrong, right?
• 5:07
  SPEAKER_00: There's a multiple different technologies and uh algorithms you can use for different purposes.
• 5:14
  SPEAKER_00: I think now where we are now the industry is starting to get a bit more advanced in their understanding of AI and the capabilities and also the limitations of the GPT models.
• 5:28
  SPEAKER_00: So I think for us, you know, uh when we meet customers now, they understand more about AI, they are curious, they really want to see how this also can be part of transforming their workflows and their industry.
• 5:43
  SPEAKER_01: You mentioned the skepticism part in the early days, and I assume that's probably lessened, but we're also in commercial real estate, which is a slow-moving beast to say the least.
• 5:54
  SPEAKER_01: So, how how do you create the trust and get around that barrier of this potentially slow uptake of things like findable and AI in this industry as comp in comparison to other industries?
• 6:10
  SPEAKER_00: Uh you it definitely I think you know real estate is not uh an early adopter of technology historically.
• 6:17
  SPEAKER_00: I think you know we we are working with uh several uh leading real estate companies, and I think within those companies you you will be able to find some early adopters and some champions that you can create some success stories around, and I think you sort of depend on that to get the rest of the pack with you.
• 6:35
  SPEAKER_00: I think when we meet customers, uh one of their main challenges is actually this you know cultural challenge of how do you enable the organization to start using AI on a bigger scale.
• 6:50
  SPEAKER_00: And I think we've had tremendous success with uh doing some early pilots with our customers with their own data, allowing them to see the results we can deliver.
• 7:00
  SPEAKER_00: Because I think for most people, and maybe particularly within real estate, their own they're not they're not capable of imagining what AI is actually capable of doing.
• 7:14
  SPEAKER_00: Because it's it's new, right?
• 7:15
  SPEAKER_00: It's a transformation and it's a paradigm shift of how we're going to work and do our job.
• 7:21
  SPEAKER_01: I think that's a really important piece, is with new technologies and um innovation and the slow generally the slow uptake, you really do need, and there's this real requirement for those early adopters and to be able to use them as case studies to enable you to push it further in the industry, right?
• 7:43
  SPEAKER_01: It's something which we found with iBoss and some of the technologies which we've implemented is if you don't have those case studies, you there's a reliance, you can there's a bit of trust in some of the early adopters, but once you got those case studies and um then the fear of missing out really starts to build, and then that adoption speeds up, right?
• 8:06
  SPEAKER_01: And so I think where are you on that journey?
• 8:09
  SPEAKER_00: It's a good question.
• 8:10
  SPEAKER_00: I think you know some of our customers I would say are early adopters of AI.
• 8:15
  SPEAKER_00: Uh some are coming around now.
• 8:18
  SPEAKER_00: I think uh not all of them have an AI strategy in place, and I think you know, uh fortunately a lot of the companies are starting to learn that you know, first and foremost, you need to get your data in order to really get the benefit from AI.
• 8:33
  SPEAKER_00: And I think you know, basically, when your data today is a wall of ring binders, it's not super easy to connect any sort of AI to that and really get the full value.
• 8:44
  SPEAKER_00: So I think it's a it's an AI journey, but it's also a digital transformation journey for some of our customers.
• 8:51
  SPEAKER_00: And uh, I think you know, in some of the areas we help our customers, we're seeing you know a 2000x improvement in some of their workflows, how efficiently they're able to do it.
• 9:04
  SPEAKER_00: And we started with the really mundane stuff of finding information that you need, you know, on a daily or weekly basis, and that's the mundane stuff.
• 9:14
  SPEAKER_00: But think about what happens when you're seeing a thousand-2000x improvement on other areas as well.
• 9:20
  SPEAKER_00: So I think this is uh, you know, either you are embracing it or I think you are in danger of being, you know, uh extinct.
• 9:30
  SPEAKER_01: Yeah, yeah, extinct.
• 9:32
  SPEAKER_01: That's uh that's a strong one, isn't it?
• 9:35
  SPEAKER_01: And that's gonna bring us on to a question which I have, which I close off.
• 9:38
  SPEAKER_01: So that would be fun to hear your thoughts on that.
• 9:40
  SPEAKER_01: In terms of in terms of that, I think we talked about off-camera, didn't we?
• 9:44
  SPEAKER_01: It's like you said, with with AI, I think where the majority of the industry is, and I think people talk about in the industry that they they're moving forward with AI, and it's quite a statement which people make, and that I think they feel like they're expected to be saying that.
• 10:02
  SPEAKER_01: Um and we're seeing, and especially at Workman, is people use utilite we're utilizing different AI tools at the moment, um, but one of the fundamental projects which we're doing is getting our foundations right and getting the data right in the first place, um, across the board, and then also digital transformation work.
• 10:22
  SPEAKER_01: So, exactly kind of what what you're saying.
• 10:24
  SPEAKER_01: And I think the industry should probably get a bit of um comfort from your comments around that, in the sense that yes, people talking about AI a lot, but actually the majority of people are trying to get their strategy in the right place and getting their foundations right at this point in time.
• 10:42
  SPEAKER_01: There's no one really that's smashing ahead and really leading the pack at this point in time that we can see.
• 10:49
  SPEAKER_01: Would you say that?
• 10:50
  SPEAKER_00: Well, I think there are some companies that are leaning more forwards than others, but I don't think we've seen a runaway uh real estate success yet within uh you know asset owners, operators.
• 11:04
  SPEAKER_00: So I think that opportunity is still uh very much there.
• 11:07
  SPEAKER_00: But again, it's a it's a question of having you know the data in order, and then you can build workflows on top of that.
• 11:14
  SPEAKER_01: And so for anyone who doesn't really know what findable is all about, one of the one of the key parts, and I'm probably s oversimplifying it, but you mentioned the ring binder walls of ring binders is OM manuals, right?
• 11:26
  SPEAKER_01: Yeah.
• 11:27
  SPEAKER_01: So operation and maintenance manuals, which tend to sit in a dark room in a ring binder on site, set up either when the building's constructed or when there's refurbishment happening, and then never sees a light of day again until a problem occurs, and then someone's got to go and try and work out what what the next steps are by going to these ring binders and hoping that they're up to date and they they've actually got the information which people require in order to do the maintenance or operation works which are required, right?
• 11:57
  SPEAKER_01: And one of your one of your strengths is basically ingesting all that data from however you're gonna get it.
• 12:05
  SPEAKER_01: So if it's from the ring binders using OCR technology, right, to make it digital, transform it, and then put it into your system, or if it's on actual digital, then bring bring that into your system as well, and then having a way of actually searching really easily to find the information which someone's looking for rather than the long drawn-out process which people tend to do at the moment, right?
• 12:29
  SPEAKER_01: Absolutely.
• 12:30
  SPEAKER_01: Have I summarized it correctly?
• 12:31
  SPEAKER_00: Yeah, no, I think that was a very good way to describe it.
• 12:34
  SPEAKER_00: Uh, but I think the only distinction I would make is that we started out with OM manuals.
• 12:39
  SPEAKER_00: Yeah, yeah, that's where we started.
• 12:42
  SPEAKER_00: And I think you know, initially we we just we were uh you know, our main task was to categorize OM uh information to be able to input into, for example, a Catholic system.
• 12:54
  SPEAKER_00: Yeah, but then we realized that you know that will allow you to navigate to find what you need, but what you really want in the in 25 is to be able to actually search or ask for the information you need.
• 13:08
  SPEAKER_00: So basically, what we built now is uh you know a GPT for your building.
• 13:12
  SPEAKER_00: So of course the chat GPT is not you know it's not trained on your building data, but our chat interface is actually trained on your building information so you can actually ask for the information you need and get it in seconds instead of hours, or in many cases, never find what you need.
• 13:31
  SPEAKER_00: I think it's you know, this documentation has been a challenge for the real estate industry for basically ever, right?
• 13:38
  SPEAKER_00: Yeah.
• 13:39
  SPEAKER_00: We have customers who have documentation on this painted roll of paper, right?
• 13:45
  SPEAKER_00: And since then we have uh you know moved into the digital age, but still the majority or the state-of-the-art document delivery in a lot of projects is a USB stick, right?
• 13:58
  SPEAKER_00: I've never seen as much use of USB sticks as in this industry.
• 14:02
  SPEAKER_00: So I'm sure the real estate sector accounts for a portion of the global usage of USB sticks.
• 14:10
  SPEAKER_00: But um, so I think, and at the end of the day, you know, having a mess on the USB stick doesn't make it easier to find what you need.
• 14:17
  SPEAKER_00: And I think many of our customers have a combination of USB sticks lying around, they have a lot of documentation on email, they have it on SharePoint, they have it on ring binders.
• 14:28
  SPEAKER_00: CDs.
• 14:28
  SPEAKER_00: Yeah.
• 14:29
  SPEAKER_00: And it's and CDs, yeah, and it's it's it's usually uh quite Celtic, it's not organized in uh in a simple way, and we usually find somewhere around 30% duplicate files as well, which also is a problem, right?
• 14:45
  SPEAKER_00: If you're trying to maintain your documentation and keep it current with the building, if you have you know five copies of your ventilation system and you're refurbishing it, and then when you're trying to replace the old documentation, maybe you're just replacing one of the copies.
• 14:58
  SPEAKER_00: Yeah, so I think the keeping the documentation up to date is absolutely essential for the assets.
• 15:05
  SPEAKER_00: I think it's always been important, but I think moving forward it's going to become even more important, as you've seen here in the UK with the you know the introduction of the Building Safety Act, which uh was absolutely needed.
• 15:19
  SPEAKER_00: Uh, and I think also moving forward with uh you know ESG, I think uh we're already seeing buildings with brown discounts.
• 15:27
  SPEAKER_00: I think this is going to just uh uh increase.
• 15:31
  SPEAKER_00: And again, if you really want to report on ESG, make the right decisions, make the right investments, uh, and also the asset value itself, I think will be close more even more closely tied to the quality of your documentation for the building because that's really the you know the that's your table of contents, right?
• 15:51
  SPEAKER_00: That's your that that's the contents of your building.
• 15:54
  SPEAKER_01: Data's gonna be the new currency in buildings, right?
• 15:56
  SPEAKER_00: Yeah.
• 15:56
  SPEAKER_01: So you're obviously very passionate about solving this problem in in real estate, um, and probably some other sectors, but why why why have you moved into commercial real estate?
• 16:08
  SPEAKER_01: Because you obviously you haven't come from that background.
• 16:11
  SPEAKER_01: What was the driver into moving this?
• 16:13
  SPEAKER_01: I think you're where were you before?
• 16:15
  SPEAKER_01: Was it uh Ambit?
• 16:16
  SPEAKER_01: Is that the right one?
• 16:17
  SPEAKER_01: Uh Ambita.
• 16:18
  SPEAKER_01: Ambita, yeah.
• 16:19
  SPEAKER_01: So you're at Ambita before, um, which was that was your first foray into like the real estate world, right?
• 16:25
  SPEAKER_01: Um doing something slightly different, but still in the technology and um commercial real estate world, right?
• 16:33
  SPEAKER_00: Yeah, well I my previous job before that was actually also within real estate, but uh but that was more within residential real estate.
• 16:40
  SPEAKER_00: So within Findable, we actually work with commercial real estate, but we also work with uh public sector.
• 16:47
  SPEAKER_00: So we have uh for example in Norway around half of our customers are actually public buildings, like uh you know, schools, kindergartens, uh everything.
• 16:56
  SPEAKER_00: And I think uh that's an area where, of course, safety is super important and also sustainability.
• 17:03
  SPEAKER_00: And I think the question of you know why real estate?
• 17:07
  SPEAKER_00: I think part of the reason why we really wanted to make an impact in real estate is because real estate is, as everyone knows, you know, represents 40% of the global climate gas emissions.
• 17:19
  SPEAKER_00: And if you really want to make an impact on that, then you know the build sector is a good place to start.
• 17:25
  SPEAKER_00: It's also the world's largest asset class, so if you want to make a big impact, it's a good place to start.
• 17:32
  SPEAKER_00: And it's an industry in, I would say, in dire need of digital transformation and AI.
• 17:38
  SPEAKER_00: And I think you know the traditional software hasn't really worked that well within real estate.
• 17:45
  SPEAKER_00: And I think part of the reason for that is because a lot of that software is just basically glorified Excel spreadsheets where you have to manually input all of the data yourself, and people in real estate aren't interested in doing all that manual job.
• 18:00
  SPEAKER_00: So I think that's where AI can come in and add some real automation and actually enable a lot of digital workflows that previously has been uh impossible to implement.
• 18:09
  SPEAKER_01: And then I I guess it aligns with BIM models were always thought it was the next big thing, but obviously very expensive and really hard to keep up to date, right?
• 18:21
  SPEAKER_01: And so this feels like a softer route in to closing out some of those problems which BIM was hoping to solve, right?
• 18:28
  SPEAKER_00: Absolutely.
• 18:29
  SPEAKER_00: And we have you know multiple customers using BIM, and I think you know, particularly in Norway, BIM is is is quite common during construction.
• 18:37
  SPEAKER_00: However, the big problem is what happens after the building is constructed.
• 18:41
  SPEAKER_00: Yeah.
• 18:41
  SPEAKER_00: And I think you know, if you're fortunate, the BIM model you get together with the building is actually what's constructed, but that's not usually the case either, right?
• 18:51
  SPEAKER_00: There are usually some adjustments, so the BIM model isn't accurate even when it's constructed.
• 18:56
  SPEAKER_00: And then, of course, the problem is how do you maintain that BIM model afterwards?
• 19:00
  SPEAKER_00: And I think that you know, in a dream scenario, you would have an always current BIM model, all documentation connected to the different objects in the BIM model, but in reality, I think we've seen maybe you know one customer who's been able to accomplish this.
• 19:16
  SPEAKER_01: Yeah, it's expensive as well, isn't it?
• 19:18
  SPEAKER_01: Yeah.
• 19:19
  SPEAKER_01: Um, so how have you found implementing this type of innovation then into this industry?
• 19:25
  SPEAKER_01: What what were the key barriers which you've you've come across and how how did you circumvent them?
• 19:30
  SPEAKER_01: How do you get around them?
• 19:32
  SPEAKER_00: I think you depend on some people who are you know forward-leaning and you know understand that if you really want to make an impact, you need to take a leap of faith in you know younger uh companies who are trying to create something new.
• 19:48
  SPEAKER_00: So I think you know, particularly here in the UK, we've been able to find some really good partners.
• 19:54
  SPEAKER_00: And now we've uh onboarded buildings such as uh 22 Bishopsgate, uh, King's Cross Estate.
• 20:00
  SPEAKER_00: We're working with Moda Living and multiple other customers.
• 20:04
  SPEAKER_00: And I think you know it's all about uh getting the ball rolling and being able to show that this can actually make a big impact.
• 20:13
  SPEAKER_00: And I think then, you know, people are getting curious and and we're getting more and more incoming uh questions and requests for for demos and prototypes, and I think that's you know, it's a it's a transformation and it takes some time, but I think uh we've been very warmly welcomed in the UK, and it's it's clear that we're solving a big pain here, and I think we're also part of the reason why we're in the UK is because we had customers reaching out to us because they had basically inspectors knocking at the door for the Building Safety Act, uh Fire Safety Act, and you know, turning around, finding a wall of ring binders, 28 days to comply, they're in a world of pain, right?
• 21:01
  SPEAKER_00: And then they call us.
• 21:02
  SPEAKER_00: So we're sort of like the Ghostbusters moving in, uh fixing it and helping them.
• 21:08
  SPEAKER_00: So, what's next?
• 21:10
  SPEAKER_00: I think for us it's uh I think first of all, you need to really have you know control over data and good structure on your data, so working a lot of that, attaching other data sources as well.
• 21:23
  SPEAKER_00: I think we really want our chat interface to be able to answer your questions.
• 21:28
  SPEAKER_00: And then I think we're also now working with uh several partners here in the UK on really understanding you know of everything we could be doing with AI, what are the main pain points we should be working on first?
• 21:42
  SPEAKER_00: And then we are you know working closely together, iterating on these problems, and then we're adding support for you know multiple workflows coming forward.
• 21:52
  SPEAKER_00: And I think you know, one of the recent uh improvements we implemented together with UK partner is uh an automated checklist for.
• 22:00
  SPEAKER_00: Functionality where you know you can take the golden thread for example and automatically our system will uh check your documentation towards the requirements.
• 22:10
  SPEAKER_00: Um we not we don't only give you a tool and an overview but we also help you fill the gaps, which is a big problem.
• 22:18
  SPEAKER_01: And then if we're looking at the wider picture with AI, kind of where where do you see it going specifically within the commercial real estate sector, um say in the next couple of years?
• 22:30
  SPEAKER_01: I don't think we can look much beyond that, can we really?
• 22:33
  SPEAKER_00: Yeah.
• 22:33
  SPEAKER_00: No, it's that's a very it's a difficult question for anyone to answer.
• 22:38
  SPEAKER_00: I think the you know people talk about AI as the you know introduction of the computer or the railway or electricity, but I think you know I've been saying this uh a lot recently that you know everybody who uses a computer for their daily work is going to have a different job already within 25 due to AI, and I think that's only going to accelerate.
• 23:03
  SPEAKER_00: I think you know, if we if we want to imagine the sort of disruption, I think we can try to imagine uh what would our jobs be like today if we didn't have a computer.
• 23:15
  SPEAKER_00: I think the paradigm we're looking at is at the same scale as that, where we're going to do everything differently.
• 23:22
  SPEAKER_00: I think the abilities for that we already see, right, with for example GPT-5, the speed of these models and the capability to ingest large amounts of data is you know mind-blowing.
• 23:38
  SPEAKER_00: And I think uh everybody should look at how they can utilize this technology to augment them to become even more uh efficient.
• 23:48
  SPEAKER_00: So I think um yeah.
• 23:52
  SPEAKER_00: Oh, it's it's difficult to imagine, but I'm uh I'm 100% convinced uh it's going to be very different.
• 23:58
  SPEAKER_01: How do you think that will affect people within the industry?
• 24:01
  SPEAKER_01: Uh do you think the industry is gonna have to evolve and change in order to accept it?
• 24:07
  SPEAKER_01: And people's rather than people losing jobs, are people's jobs gonna change to align with newer processes, better better approach with AI?
• 24:19
  SPEAKER_00: I think right where we are right now, I think it's uh you know it's definitely I wouldn't be too concerned about uh that that you know losing a job.
• 24:29
  SPEAKER_00: I think it's more a question about uh using the right tools and being augmented where you know our customers are finding that AI is like giving them superpowers, and particularly over maybe some of the most mundane tasks that nobody really wants to do.
• 24:49
  SPEAKER_00: We've organized now and processed over 30 million pages of real estate documentation, and you can imagine how many manual hours of work that is, and I think we've met a handful of people uh during the last five years who actually enjoy sitting there working, organizing, categorizing documentation.
• 25:09
  SPEAKER_00: So I think that's just one example of you know mundane tasks that AI is going to handle for us.
• 25:15
  SPEAKER_00: Another example is you know meeting notes.
• 25:18
  SPEAKER_00: Who uh is writing meeting notes anymore, right?
• 25:21
  SPEAKER_00: The AI is going to do that for us.
• 25:24
  SPEAKER_00: The AI is automatically going to process your meeting notes for the previous meetings and help you prepare and you know produce better meetings.
• 25:34
  SPEAKER_00: So I think the the most mundane tasks is going to be removed, and then I think people will be spending more time on high-value work.
• 25:45
  SPEAKER_00: So it's like everybody has a super advanced assistant, right?
• 25:49
  SPEAKER_00: That it's not able to do 100% of the job, but they can do 80% of the job, and then uh we can use our human brains on the most important 20%, and I guess getting people to do what they they love doing rather than the stuff which they absolutely love with doing, right?
• 26:06
  SPEAKER_00: Yeah, I think we have you know uh multiple customers who uh utilise Findable because they had an ambition to I mean, they're actively working to remove the most boring jobs for their employees, right?
• 26:21
  SPEAKER_00: So it's it's not just about becoming more efficient, but it's also about uh employee satisfaction.
• 26:27
  SPEAKER_00: So I think, and also what surprises us with some of our customers is that the data which until now has been locked inside of these ring binders or USB sticks in the hand of few people who know where to find stuff is actually very valuable for a large portion of the organization.
• 26:45
  SPEAKER_00: And being able to just ask questions and get the answers you need is super valuable for other people and not just the facility manager.
• 26:54
  SPEAKER_01: Yeah, and allows you to give a better level of service to your tenants, to to your landlords, to everyone, really.
• 27:01
  SPEAKER_00: Absolutely.
• 27:01
  SPEAKER_00: And it has an impact on your uh you know, your asset value, it has a big impact on your ability to certify your building, on your ESG reporting.
• 27:11
  SPEAKER_00: It all its data is really connected to uh to how to operate uh your assets, right?
• 27:17
  SPEAKER_00: On the value of your assets.
• 27:19
  SPEAKER_01: Yeah.
• 27:20
  SPEAKER_01: And going back to the golden thread side of things, can you just give a bit more detail on maybe case study in terms of where you've where you've really helped out someone in the UK with that side of things because that's quite a scary topic for people and quite uh quite a topic of where where do we start really?
• 27:38
  SPEAKER_01: Um so it'd be great to understand how how you guys have assisted with that.
• 27:42
  SPEAKER_00: Yeah, I think some of the buildings we're working with, you know, have I think we have uh at least I think we have four buildings in our platform now with more than 40,000 documents per building.
• 27:54
  SPEAKER_00: So you can imagine the sea of information that you're trying to you know sift through to find what you need.
• 28:01
  SPEAKER_00: And I think for the uh Golden Thread is about organizing information, making it easily digitally available.
• 28:10
  SPEAKER_00: And I think you know, without that foundation, it's very difficult to comply with the Building Safety Act and particularly the Golden Thread.
• 28:19
  SPEAKER_00: And I think uh you end up spending a lot of money on consultants to manually sift through your ringbinders to find information.
• 28:28
  SPEAKER_00: So I think that you know, for us, we really care about enabling safe and sustainable buildings.
• 28:36
  SPEAKER_00: So I think that's why it makes a ton of sense for us to be able to support uh the Building Safety Act, the Golden Thread, and uh and that gives us a lot of uh purpose.
• 28:47
  SPEAKER_01: I guess one question which I have about how the technology works is say if you've got your original OM manual and then a refurbishment happens and you're pushing that information into your technology.
• 29:00
  SPEAKER_01: How does it distinguish?
• 29:01
  SPEAKER_01: How does it if you ask a question of say, what's my VRF system, if that's changed in refurbishment, how does how does it respond?
• 29:08
  SPEAKER_01: How does it know?
• 29:09
  SPEAKER_00: Yeah, well, that's a very good question.
• 29:11
  SPEAKER_00: I think you know, right now the fact that we're categorizing the documentation makes it very it makes it a lot easier to replace the old documentation.
• 29:22
  SPEAKER_00: So instead of having you know, the the worst example I've seen is a building here in the UK where they had a couple of thousand documents.
• 29:31
  SPEAKER_00: The file names were all numbers.
• 29:34
  SPEAKER_00: And then when you have no way of searching the contents of those files, you know it's super difficult to find anything and replace the outdated documentation.
• 29:42
  SPEAKER_00: So uh one thing we do is that we have implemented an automatic file name for example.
• 29:48
  SPEAKER_00: So we use AI and the contents of the documentation to suggest file names for these thousands of number named documents, which allows you to find and replace.
• 30:02
  SPEAKER_00: We also, however, um for us dokumentation is you know, the documents are basically numbers.
• 30:10
  SPEAKER_00: And we can actually uh mathematically measure the distance between the documents in uh you know in a three-dimensional vector space so we can see, you know, we can show you uh we can show you exact duplicates but we can also show you similar documentation, uh semantic similarity between uh documents.
• 30:34
  SPEAKER_00: And this uh will allow us to also suggest documentations that you might want to be archiving based on the similarity with what you're uploading.
• 30:46
  SPEAKER_00: So I think that you know we really want to enable our customers to have up-to-date uh documentation which is in line with what's actually in the building, yeah.
• 30:57
  SPEAKER_00: Uh, and we're working hard to accomplish that.
• 31:00
  SPEAKER_01: So setting up Findable, being a founder, how have you found that?
• 31:05
  SPEAKER_01: Um and how have you found being a startup?
• 31:09
  SPEAKER_00: Oh, it's certainly a roller coaster ride, I have to admit that.
• 31:13
  SPEAKER_00: Uh it's you know it's uh you have to deal with everything, right?
• 31:20
  SPEAKER_00: And there's a lot of ups and downs, but I think at the end of the day, I'm super proud of uh what we built and uh the team that we have assembled.
• 31:31
  SPEAKER_00: I think you know we have a fantastic team of highly motivated, dedicated, uh exceptional people who are working hard to make this a reality, and I think seeing that interaction between our team, our technology, our customers, and what we're able to achieve together is truly uh amazing.
• 31:51
  SPEAKER_00: So I think you know I uh I mean I don't understand why everyone don't want to become a founder, but at the same time, it's you know it's certainly challenging.
• 32:03
  SPEAKER_00: Trying to give the balance between the family life and being a founder, for example, is uh certainly a challenging aspect.
• 32:09
  SPEAKER_01: Yeah, um I know that feeling.
• 32:12
  SPEAKER_01: It's uh sacrificing probably a lot of sleep.
• 32:14
  SPEAKER_00: I mean it feels like uh I should have deserved to get a couple of hours of extra every day, but you know, it's 24 hours for everyone.
• 32:22
  SPEAKER_00: Yeah.
• 32:22
  SPEAKER_00: And I try to remind myself that you know even Bill Gates has 24 hours each day.
• 32:27
  SPEAKER_00: So we're on a level playing field when it comes to the number of hours.
• 32:31
  SPEAKER_01: So you mention you mentioned a family what's what what's your family?
• 32:35
  SPEAKER_01: They're obviously based back in Norway as well.
• 32:37
  SPEAKER_00: Yes.
• 32:37
  SPEAKER_00: I have uh two sons, five and seven.
• 32:40
  SPEAKER_00: Uh so it's we have a really active family, and trying to balance that is certainly a challenge, but that's like uh you know, it's the joy of my life, and being able to combine that with something which is you know creating a company is uh I feel very fortunate.
• 32:59
  SPEAKER_01: And so where are you in terms of startup wise with with Fundable?
• 33:03
  SPEAKER_01: Have you what sort of numbers are you at in terms of employees?
• 33:06
  SPEAKER_01: And then also have you gone through VCP to get to where you are?
• 33:11
  SPEAKER_00: Yeah, we've we've been fortunate to get you know several investors to support us along the journey.
• 33:17
  SPEAKER_00: And we last summer we finalized our Series A round with venture capital from 0.9, which is one of the leading uh SaaS VCs in Europe.
• 33:28
  SPEAKER_00: Uh, and uh our team now is uh around 30 employees.
• 33:33
  SPEAKER_00: We have a team here in London and a team in uh in Oslo.
• 33:37
  SPEAKER_00: I don't think we have an ambition to grow an enormous team.
• 33:40
  SPEAKER_00: I think you see that a lot of the uh AI companies that are really scaling these days, like Lovable, for example, are able to do a lot with a relatively small team.
• 33:50
  SPEAKER_00: So I think that is, you know, we don't have a goal in itself to become a massive team, but uh uh you know we're definitely using you know AI ourselves in how we do our tasks every day.
• 34:02
  SPEAKER_00: Uh, all of our developers are of course using uh AI in their development tools.
• 34:06
  SPEAKER_00: I think the level of efficiency we're able to find through using the right tools is truly amazing.
• 34:13
  SPEAKER_01: It's it's interesting.
• 34:14
  SPEAKER_01: I think it's that's I'm I forget who um who's mentioned it, but we're we're close to seeing probably that that first one one person, one billion pound company, right?
• 34:24
  SPEAKER_01: Yeah.
• 34:24
  SPEAKER_01: Um and when I don't think we're very far away from that.
• 34:29
  SPEAKER_01: We we mentioned the VC thing, and I think that's a whole world which commercial real estate people, unless you're in the prop tech world, really have any exposure to or the understanding.
• 34:40
  SPEAKER_01: It'd be great just kind of you to explain that journey because I keep on what using the word journey, so I'd like to cut that out and remove that because I think I've said journey about 1500,000 times.
• 34:52
  SPEAKER_01: Um it is a journalism, but it is a journalism, it is a journey.
• 34:55
  SPEAKER_01: Um maybe we keep it.
• 34:56
  SPEAKER_01: I don't know.
• 34:57
  SPEAKER_01: We just keep this whole monologue of me killing myself.
• 35:01
  SPEAKER_01: Um but the VC process feels very foreign to people in commercial real estate and probably have will never have any exposure to it unless they move into the prop tech world or something like that.
• 35:14
  SPEAKER_01: So if you wouldn't mind just explaining what that is and the fun and games that you have with doing that.
• 35:22
  SPEAKER_00: Absolutely.
• 35:22
  SPEAKER_00: I think you know the the VC and the funding world is its own strategy and dynamic.
• 35:29
  SPEAKER_00: And I think uh, you know, VC, also venture capital, is as uh you know it's a necessary part, I think, of creating the sort of companies that are changing uh the world and particularly tech companies.
• 35:45
  SPEAKER_00: I think that our journey started getting some early support from some forward-leaning companies in Norway and startup accelerators.
• 36:00
  SPEAKER_00: And then we moved on to getting our first international VC on board, a company called Kumpas out of primarily out of Denmark, who invests in sustainability and the built-world in our seed round.
• 36:14
  SPEAKER_00: So we did a pre-seed round early on, then we did our seed round, which was still also very early, and they took a chance on us, which is you know part of what VCs have to do.
• 36:25
  SPEAKER_00: And then we did our Series A round together with the point nine in Lead.
• 36:31
  SPEAKER_00: And I think you know, to develop this kind of solutions, it's a bit of a chicken and egg problem, right?
• 36:40
  SPEAKER_00: Because if you're not able to build something, then customers won't buy it.
• 36:46
  SPEAKER_00: So basically, very few customers buy it before you build it.
• 36:51
  SPEAKER_00: So you have to make sure that you build the right thing, and in order to get there, you need capital.
• 36:57
  SPEAKER_00: And uh the efficient way of doing that, because most likely we'll be unable to attract like a bank loan.
• 37:05
  SPEAKER_00: Uh, so the people who you know understands what you can build and what's the potential of changing, you know, disrupting markets to create uh tremendous successes are typically VCs.
• 37:19
  SPEAKER_00: Uh, so they they've taken a chance on us, uh, they believed in what we were building, they've uh funneled you know uh uh multiple millions of euros into this and made that investment to a certain degree on behalf of real estate, right?
• 37:37
  SPEAKER_00: And our customers.
• 37:38
  SPEAKER_00: So I think that uh you know it would be like approaching a customer and asking, you know, we're gonna spend 10 million euros on solving one problem for you.
• 37:49
  SPEAKER_00: What problem is that?
• 37:51
  SPEAKER_00: I think very few companies are able to mobilize that sort of effort into solving a problem uh operationally.
• 38:01
  SPEAKER_01: And how how involved are the VCs once you've once you've received funding from them?
• 38:08
  SPEAKER_01: Do you get some of their insight, their help, and their network, or how does that work?
• 38:14
  SPEAKER_00: Absolutely.
• 38:15
  SPEAKER_00: So I think it depends very much on who you know work with and partner with, and I think you know that's a basically a 10-year marriage, right?
• 38:24
  SPEAKER_00: So I think uh what you quickly learn that is that it's all about people and you know who are the actual partners you are working with.
• 38:34
  SPEAKER_00: Uh we're very fortunate to have two excellent partners we're working with: Sebastian and Kompas and Christoph in 0.9.
• 38:42
  SPEAKER_00: And I think you know they they're great supporters of us, uh, but they also recognize that you know we need to build the company.
• 38:52
  SPEAKER_00: They're not gonna build it for us, but they have been a part of investing in you know hundreds of startups before us.
• 38:59
  SPEAKER_00: So they know some patterns on what's work, what works, and what doesn't work.
• 39:04
  SPEAKER_00: And I think a major challenge uh when you're starting a one of the biggest challenges of building a company, I think, is you know attracting the right people and the right talent.
• 39:16
  SPEAKER_00: And I think that's an area where uh VC is super helpful because they've seen it so many times before, they see they know what works and they know what doesn't work.
• 39:28
  SPEAKER_00: So working on our strategy, working on our talent is been uh they've been uh a really important contributor.
• 39:35
  SPEAKER_01: And then I'd I think I'd really like to hear obviously what that kind of that light bulb moment was for you in terms of you're at MBT and you're working there.
• 39:48
  SPEAKER_01: What was that light bulb moment for you where you gone findable?
• 39:53
  SPEAKER_00: Yeah, it's I don't know if there was you know one major light bulb moment, but I think you know, to be honest, uh it was also very much connected to uh to Covid because I think in Ambita we were closely connected to residential real estate, which completely stopped from one day to the other.
• 40:14
  SPEAKER_00: So the company had to make some tough decisions.
• 40:17
  SPEAKER_00: Uh, and uh I decided to part ways with uh Ambita in that process, and I think that it was you know in that dynamic that you know all of a sudden Lars, Knut, and me were you know uh out on the market trying to figure out what they were going to do, and at the same time, we had experienced this amazing capabilities of these large language models, this uh AI algorithms on documentation and the capabilities there.
• 40:48
  SPEAKER_00: So I think we uh we really wanted to see if there were other you know areas where we could utilize this technology because I think for us there was there was no question of whether or not AI is going to change how we work with documents.
• 41:04
  SPEAKER_00: Uh, and you know, we we wanted to see if we could be a part of that transformation.
• 41:09
  SPEAKER_00: And initially, for the first couple of months, our hypothesis was actually to work with uh legal tech and lawyers.
• 41:17
  SPEAKER_00: Uh, and uh we looked into that space, but I think uh back then you know the main problem was that you know lawyers bill by the hour, and they didn't necessarily want to become more efficient because that would mean that they could bill for fewer hours.
• 41:34
  SPEAKER_00: And at the same time, you know, we came back into real estate, and uh I was really surprised when I learned that 40% of climate gas emissions are coming from real estate, and that sort of led us in that direction.
• 41:50
  SPEAKER_00: And the more people we engaged with, the more clear it became how massive the document problem is within real estate.
• 41:58
  SPEAKER_00: And you know, the 30 million pages we've organized this far is just scratching the surface, right?
• 42:04
  SPEAKER_00: So I think that it's it's maybe one of the industries with the most documents.
• 42:11
  SPEAKER_00: Uh, and I think also another aspect of real estate is that you know a building in Norway isn't terribly different from a building in the UK.
• 42:19
  SPEAKER_00: Yeah, uh, but if you're building something for Norwegian law, it's completely irrelevant in the UK.
• 42:25
  SPEAKER_00: So I think for many of our customers are do you know have international portfolios of assets.
• 42:32
  SPEAKER_00: So, you know, one of our customers is uh Verlux out of Denmark, who has you know 80% of the roof window market, they have buildings in 36 countries globally, and they're using us as their global document platform.
• 42:46
  SPEAKER_00: And you know, buildings in countries are mostly similar, and I think that's uh also an advantage working in this industry.
• 42:54
  SPEAKER_01: Yeah, and also from dealing in a in a world where we've we've been touching different countries and different languages, being able to be in your own country and speaking your own language and being able to search on drawings and finding documents in a different language where it's it's not being translated is is hugely helpful.
• 43:14
  SPEAKER_00: Yeah, absolutely.
• 43:15
  SPEAKER_00: And I think you know, when we processed our very first building, uh I think we were uh you know genuinely surprised to find because we were expecting Norwegian documents, right?
• 43:26
  SPEAKER_00: But I think we found 10 different uh languages in those documentations, and that was a couple of thousand files for one building.
• 43:33
  SPEAKER_00: So I think from the very first building in 2020, we realized that we need to build uh multilingual uh multilingual support in our platform.
• 43:44
  SPEAKER_00: And uh one of my co-founders, Knut, he used to work at Google, and one of the projects he worked on was the internationalisation of Google News globally.
• 43:54
  SPEAKER_00: So I think we have, you know, ever since the very first building, we've built language support into everything.
• 44:01
  SPEAKER_00: So I think that's also one of the reasons why, for example, the elux uh find this uh so transformational is because they can be sitting in Denmark asking questions about the Spanish building and getting their answers in Danish.
• 44:16
  SPEAKER_00: Yeah.
• 44:17
  SPEAKER_00: And it it doesn't matter what the language the original documentation is in, you can get the answers you need.
• 44:24
  SPEAKER_01: Yeah, and I guess really helpful in in the wor in somewhere like the UK and specifically London as well, where you have a lot of people who actually their first language might not actually be English as well.
• 44:35
  SPEAKER_01: So being able to search in their own native language as well whilst working on English buildings will be really, really powerful.
• 44:42
  SPEAKER_01: Closing up, big picture.
• 44:45
  SPEAKER_01: Um you've you've alluded to it, but I think um if you get to understand just very quickly why is change required in commercial real estate and why is it so important?
• 44:57
  SPEAKER_00: I think change is coming to every industry, whether you want it or not.
• 45:02
  SPEAKER_00: I think uh you know, I in some of my presentations I usually have this image of a tidal wave with AI.
• 45:09
  SPEAKER_00: I think you know it's going to hit everyone, whether you want it or not.
• 45:14
  SPEAKER_00: And I think, you know, uh at the moment, I think for some it's an opportunity to have as a competitive advantage leaning in.
• 45:23
  SPEAKER_00: I think in not long, I think everybody will need to utilize it to stay competitive.
• 45:29
  SPEAKER_00: I think it's going to open up fantastic opportunities, also, particularly within uh safety and sustainability, uh, and of course profitability.
• 45:38
  SPEAKER_00: So I think that uh AI is going to change the real estate industry for the better, uh, and really we really want to be a part of that.
• 45:46
  SPEAKER_01: And then finally, um a question which we close on on every episode is does commercial real estate require evolution or revolution?
• 45:56
  SPEAKER_01: And that's going back to my point, which we talked.
• 45:59
  SPEAKER_01: You you sort of alluded to some of it, and I I think it'd be great to hear your view on that.
• 46:05
  SPEAKER_00: I think for some it's an evolution, and for some it will be a revolution.
• 46:09
  SPEAKER_00: So I guess it depends a bit on you know, are you leaning in or not?
• 46:13
  SPEAKER_00: And I think you know, we have some customers have been working with AI for years already, and we have some who are you know just starting to open their eyes.
• 46:22
  SPEAKER_00: So I think for for many in real estate it's going to come as you know uh surprise.
• 46:29
  SPEAKER_00: But I think for most, I think we see people leaning in, and I think you know, as we talked about before we started, uh you know, we know already see uh AI transforming old industries like for example legal tech and health tech are embracing AI and the opportunities within AI.
• 46:50
  SPEAKER_00: I think you know, uh of course, in the future when you visit your doctor, he's not going to spend you know his uh focus on typing on the computer when you try to explain, you know, uh why you're there for the appointment.
• 47:05
  SPEAKER_00: And of course, he will have access to you know the vast amount of knowledge of all medical science at his fingertips or even maybe automatically, right?
• 47:15
  SPEAKER_00: I mean uh so I think it's going to transform every industry and it's going to transform real estate.
• 47:24
  SPEAKER_00: And I think you know we really just uh encourage our uh customers and potential customers to get started trying stuff and be open-minded.
• 47:35
  SPEAKER_00: And uh I think you know many of our customers have tried to implement, for example, uh Microsoft Copilot and haven't really succeeded with that across you know a large amount of documents and buildings, and for us then to come in and show them you know what it actually can be like is uh is great.
• 47:54
  SPEAKER_01: Thank you so much for joining us today on uh Built to Perform.
• 47:56
  SPEAKER_01: Uh it's been a really interesting conversation, and um, it's great to see your passion for commercial real estate, how to change it specifically utilising AI.
• 48:07
  SPEAKER_01: Um, and I think maybe one of the ending thoughts is probably embrace it or become extinct, in in your words.
• 48:17
  SPEAKER_01: Yeah, I would recommend embracing it.
• 48:20
  SPEAKER_01: Well, thank you again.
• 48:21
  SPEAKER_01: Um, it's been great having you on.
• 48:23
  SPEAKER_01: Thank you very much.
• 48:24
  SPEAKER_00: Thank you for having me.