GP 56: Get Ahead With AI and How To Double Down: with John Williams of Scail UK

In this episode of the Grow Places Podcast, Tom Larsson is joined by John Williams of Scail UK to explore what it really means to get ahead with AI — and where the built environment should be doubling down as technology accelerates. John shares his journey from journalism, through senior marketing roles in property, to building AI-led businesses focused not on hype, but on unlocking the story inside organisational data.

Together, they discuss why AI is only as powerful as the data and context behind it, why real estate still struggles to understand the “why” behind customer decisions, and how unstructured conversations — from stakeholder meetings to emails — can become structured intelligence when approached with the right methodology. The conversation expands into the future of cities, personal AI agents, and what it might mean for buildings and neighbourhoods to publish meaningful data rather than relying on static digital interfaces. At its heart, this is an optimistic discussion about AI removing administrative burden and amplifying the most human parts of our work: judgement, creativity, synthesis, and relationships.

Transcript

Meet John: Journalism To Real Estate

SPEAKER_00 0:02

Hello and welcome to the Grow Places Podcast, where we explore the virtuous circle of people, growth, and place. Brought to you by Grow Places and hosted by our founder, Tom Larson.

SPEAKER_02 0:15

John, thanks very much for joining me today on the Grow Places Podcast. How are you? I'm very good.

SPEAKER_01 0:20

It's a little bit damp in London, but I think the rain has stopped.

SPEAKER_02 0:23

Exactly, exactly. And and we first met at uh Joy Nazari's Place Camp. So shout out to Joy, who was a past guest on the podcast as well. And I was just fascinated by your talk, your journey. And so I think it'd be really good to start at the beginning.

SPEAKER_01 0:38

Journalism, was it for you? It was, but before we we get into that, I would give a shout out to Joy because I thought Place Camp was one of the best run days and corporate events I've attended in a long time. So big uh big kudos to her and her team for that because it was awesome, wasn't it? It was really good. It introduced people like you and I because there was a nice collegiate atmosphere to it, which was uh yeah.

SPEAKER_02 0:59

And and what I really enjoyed being in the real estate industry is not hearing from real estate people, um, with all respect to real estate people. It's nice to hear from people who are peripheral or um associated with it, and I think she curated a really nice mix for that. So um, yeah, it'd be good to hear about your journey,

Planning Meetings And Data Mindset

SPEAKER_02 1:15

your place in real estate, how you see yourself as well.

SPEAKER_01 1:17

Um sorry, I'll try and give you the short version, but um, in essence, I started as a journalist, um, trained as a journalist with Trinity Mirror, so the mirror training scheme. And within the realms of that, I attended, oh gosh, probably 500, 600 planning meetings back in the day, as you do when you're a trainee journalist. And so as a young Cub reporter, I got a real exposure to the, shall we say, the detail, the kind of ennui of the planning process, its limitations, the fun elements in a way most 20 somethings wouldn't. So when I left journalism, moved into comms and then digital marketing, I had that kind of grounding in that world of seeing how local and regional councils drive infrastructure and local planning and strategic frameworks in a way that few people do, unless you work in the industry.

From Comms To AI Entrepreneurship

SPEAKER_01 2:07

Um, that ended up with me um working at Night Frank and then subsequently the instant group as CMO. And the instant group, for those of you who don't know, with the kind of the flexible workspace specialists. I've started an AI business, which is not so much about the AI, it's about the story behind your data. So the data that sits in every business or every organization that people aren't really cognizant of or being data, or that can be organized as data to give you amazing outputs. And that's uh when you can unlock the power of AI on a proprietary basis, is to think right, what amounts to our business, what data sits within it, or what can we add into our data structure, and then apply AI to drive some really interesting results?

SPEAKER_02 2:53

Yeah, amazing. And um, so so why do you think if you were to sort of contextualize your first role as a journalist and your role now as an A to I business, what are the similarities and differences between how you approach

Facts, Narrative, And Marketing

SPEAKER_02 3:08

that?

SPEAKER_01 3:08

Well, I I think fundamentally the actual the genesis of it is quite similar because they're both about storytelling in in a way that placemaking is to a large degree as well. Journalism is about understanding the facts, the the underlying story, and then adding a narrative to it, making sure that you've reported both sides and being as mechanical as you can about that, even in a creative industry. And what I've learnt about marketing is that it is founded on data, and it's founded on that idea of what are the building blocks that allow me to form a narrative? And in short, AI runs off a very similar premise. So data structure or even unstructured data is what can we have that represents the building blocks for our business in data form? What's the story you want to tell? And you'll find that AI will tell that story for you if you've got those foundations in the right place.

SPEAKER_02 4:04

Yeah, yeah, amazing. I really love this framing of facts and narrative. Um, because that's how we think about our business at Grow Places. It is on one hand, you know, we want to be led by data, and on the other hand, we're out talking to people, we're doing these podcasts to communicate what are complicated themes in a more human way. Yeah. And and uh we sort of think that that is almost the future of business and humans' role within society more broadly, if you want to go that big, is the narrative side uh overlaying with the facts and being able to harness both. You know, what do you think about that?

SPEAKER_01 4:36

Well, I I think there's the you know, even I working as I do in an adjacent area, there is that perceived threat of AI being reductive to humans. And that is slightly inevitable. Yeah, I think even thinking back to the internet before, well, computers before that, there is that reductive threat that this will remove the need for us. But what I'm seeing is just more and more use cases for for bringing humanity into

AI As Optimism: Humans In The Loop

SPEAKER_01 5:01

business problems and removing the more robotic elements of our roles. So I'll give you an example, you know, analysts focusing on the gathering of data and the sorting of data, that kind of role or that function that is going to be removed because the AI will do that for us. But you need to have in your mind, the analyst needs to have in their mind that kind of narrative, right? Where do I need to pull in data? Is it accurate? What do I apply to it to create a better whole, in short? And to your point around the development world, that's a similar story, I would argue. It's that the rhythm rule will be removed, leading you to think strategically and creatively. And strategically and creatively, not just about things like marketing materials, vision statements, and so on, but what can we bring in to really stand this scheme up? Where can we drive margin? You'll think much more laterally about those things and then apply AI to those problems to add to it and add volume.

SPEAKER_02 5:58

Yeah, yeah, no, absolutely. And um, you know, we always we we think about, as you say, if you can remove some of that administrative work, we have more time to spend visiting places or thinking about places or talking to communities, and as you say, doing the human bit of the process um in a better way and and then sort of feeding those inputs back into the company brain, however you want to call that, which is part human, which is part digital. Um, and I think that's really enriching and really optimistic, I must admit. Like um, I've very much sit on the optimistic

Turning Conversations Into Structured Data

SPEAKER_02 6:27

side of what this transition is bringing forward.

SPEAKER_01 6:30

Um But with that just to interject for a second, that's that's one use case I find really fascinating and and and fun around the use of AI, is if you're talking to people to take expert views, you know, that would have been something you'd have to spend a long time note-taking, synthesizing, reproducing. And I think what we have now is this the fact that you can record everything, you can transcribe it, you can then introduce elements of automation if you store that information the right way, and if you add context to it. And we can come to that later in our in our chat. But those unstructured conversations can sometimes sometimes now become very structured data points to prove business cases, to compare side by side if you break them down into their constituent parts. So in a simple discussion with a local stakeholder, you take a view, you want to compare it to other views to give you a scientific reason approach. AI would do that really well. But you as the human need to add context understanding so they become comparable.

SPEAKER_02 7:31

Yeah. Does that make sense? Yeah, yeah, totally, totally. And um, I think back, you know, as a as a journalist, what in in some sense you're those facts, you're you're capturing real-world, real-time data, aren't you, in some respects? And then you're telling a story about that. And I do think we can be really optimistic about the the inputs to this process, the amount of data that we we haven't even scratched the surface on what data we can um harness, whether it's real-world data or or um, you know, biological data, emotional data. I just think there's a there's a huge amount more we can learn about ourselves and the places we operate in. Um, and so so I wonder, sort of on that journey, how where do you think we are in terms of today, um, with obviously LLMs and um diffusion models essentially as the two primary aspects of of AI and and and how how do you shape that in your work today and and maybe looking forward because there is also an element of you know, how do we go from zero to one, isn't there, in all of this? Like what steps do we need to take today? There's a lot of questions, Tom. Sorry, often there are. My questions are often very rambling and big picture. You can you can pick up on whatever aspect you want.

SPEAKER_01 8:42

Um I'll try

What Counts As Data In Property

SPEAKER_01 8:43

and uh answer the first bit first. So so what constitutes data, and that that is a a conceptual leap I think most businesses need to take. And I don't want to be overly damning of real estate or commercial or commercial property in particular, but there is a sense that data is organized and structured in Sits and Excel or databases, then it is something you'd you'd identify as a as looking at feeling like data. You allude to a more interesting point that is actually there is so many different sources of data, and it's how you apply a methodology to them. So we had the initial thing around um uh speaking to local people and trying to draw information, inference from that. Now, clearly it's more about how you develop a methodology around those conversations. How many people, in what areas, how do you align them, which postcodes, if you are you need to organize them, then you structure that in a database, then you apply AI. And then AI can draw inference and understanding from that, if that if that's clear. And I think to answer the second part of your question, we are miles off in terms of really harnessing quite what that looks like. And then from a sector, I mean, there is so much to play for because there probably hasn't enough work been done yet around really structured surveys at scale to understand the why of how we do things. There's still a lot of opinion and conjecture in them, and clearly a lot of work too. But Joy's place place camp was an interesting one because the talks were there's still a lot of data of information and opinion around why the thing was done, why it was successful, and no one stood up really and said, Well, actually, we we mine 5,000 different stakeholders, which is

We’re Miles Off: Sector Gaps

SPEAKER_01 10:23

what you can do with AI. Yeah, and we then we got we synthesised their views, we produced a huge report based on that, and that would have been that will be push button shortly. You will be just able to go to people at scale, derive views, and then look at them side by side to make bigger decisions, yeah. And that's the exciting potential, I think, uh, to answer your uh overlong question with an overlong answer. I think that's the potential is when people free themselves up from the rigmarole as we discussed it, to thinking, actually, big picture, what do we need to harness in our business to make much bigger, better decisions and be bolder? And that's when it will get interesting. I think.

SPEAKER_02 11:00

Yeah. Yeah. So how do you how do you begin that process then? If you're working with a client or an entity, where where do you start? Where's okay?

SPEAKER_01 11:10

So the the number one use case that I'm finding for for now, going into businesses and trying to understand what their biggest issue are. This is the most simple use case. Most businesses don't have a good enough view of their customer. And so nearly nearly everyone is using CRMs of some shape or other. But it's they're usually underutilized, they're over-reliant upon human um note-taking ability or human input. So, really, the level of data stored in CRMs is is suboptimal because you can't really make great decisions from it. I'm sure you have seen the likes of Salesforce and HubSpot are all plugging in AI and they're Air Force. They do Agate Force and they've got Breeze and HubSpot. They're really good tools. What people don't read in the small print is this is highly dependent on the quality of the data you're inputting and how much data. And again, and I don't want to over-generalize, but in property in particular, we haven't been good at harnessing and capturing customer data seriously.

Start Here: Fix The CRM

SPEAKER_01 12:06

And therefore, that is going to hinder us from the get-go in our adoption and use of AI. So, to go back to my point, that primary use case is why do we need a CRM? How are we going to enhance it? What do we put in there? And thinking progressively about that. We need to capture more opinion from our customers. Why do they buy? What do they want, or make them buy again? And yeah, Todd, I'd be interested to get your view on this. I think in other, should we say more consumer direct sectors where you can automate these answers, or in big B2B, like Dell or Microsoft, they're really good at finding out information about their customers. It's it's harsh, but I don't think real estate is actually very good. It's there's still a lot of supposition and we don't know enough about the why people do things. Would you agree?

SPEAKER_02 12:52

Yeah, yeah, totally. And um I think that's absolutely fair. And and I also think that because often in property as well, um any business's direct customer isn't often the end user. And so um often this may even be finding out what other businesses want and need, and I don't think we necessarily do that as well as we could. And and then the other thing which gets thrown around, I'd be interested to get your thoughts on it, is that you know, real estate is a people business in inverter commas, so we're kind of going to be okay because it's gonna just carry on as it is. So I don't know how you would contextualise that sort of statement.

SPEAKER_01 13:28

That that's a it's a big question. So I might go back to the point around data first of all. Because I think once you've sorted out that base case for data within the business, which is who you know, who is your customer, why do they buy, what data points do we have on them? Because really you want double figures, you want 10 or 12 data points to make good decisions and hand it over to an agent force or a breeze. One. But then two, when you know that, what are all the other potential sources you can overlay against it? Because when you have organized data in a structure which is understandable and intelligible to AI, that gives you the power, and this is where I think it gets really exciting to bring in other data sets. And in property, you know, we're pretty good for open source uh data sets, things like land reg data, uh completions, planning portals, all these different areas. But in other sectors, this is going to become big news. What third-party data sources do you pull in

Enriching With Third‑Party Datasets

SPEAKER_01 14:23

to enrich and enhance your base data set or your planning rationale or your development thesis and strategy? What do you wrap around that? And again, I think this will be an interesting challenge for our sector because a lot of people think about third-party data sets and they analyze it and look at it side by side. AI will enhance that rapidly, pull it together, and allow you to just crunch it together to look at patterns and um correlation between the two or three or four different data sets in a way you would have need a big data analyst team. That will become more accessible, but again, it's the power of your imagination. What do we need to bring in to make that stand up? Yeah. And that so that was your your first question. The second was around people, and and we're we're we're a people to people business. Yes, but when I'm talking about automation in this scale, automation means being more qualified to go into a meeting. So, by example, more research on the person you meet, the type of things they like, the the again, the view of them in the data world. And I used to talk about this back in the social media SEO days, is you should think of yourself, you will think of yourself as a personal brand through your podcast, through your LinkedIn profile, and your website. And those are all data points that the LLMs will be picking up and going, so Tom is this guy, he does this through these mediums, and he is an expert in these areas. And that's a lot of data points for an LLM to make a decision on. Someone who isn't Tom, who hasn't updated their LinkedIn, doesn't do much on LinkedIn, doesn't have much profile, where is the LLM basing decisions on you? Right? So it's a simple kind of what what data source do they pull from you? That's that personal branding discussion will become increasingly interesting. Because there will be a you and there's where there will be a digital you.

Personal Brand As Data Signal

SPEAKER_01 16:11

And a bit of that's social media, a bit of that's Google and so on, that's all that's been there for 10, 15 years, LLMs will be making bigger decisions about you and your higher ability and your your deal making ability and so on and your uh success criteria. Is that a match to your client base?

SPEAKER_02 16:26

Yeah, yeah. And I'm I'm sure some people listening to this would go, oh well, uh I'd rather not be known to the LLMs then. So I'd rather stay off these platforms. How how do you think about that kind of Yeah, yeah. That that that that profile point um and and and to to the value that people are gonna have going forwards.

SPEAKER_01 16:46

Well, I'd say if you want to stay off those pro those platforms, good luck. I mean, that they're gonna make a decision based in one way or another. This is I used to have this discussion point with with clients back in the day about, oh, uh I've uploaded a video to YouTube, it's not very flash, or I did this when I was a student, please can we remove it? And it's gonna exist somewhere and somehow. So, really, how do you mitigate it? And the LLMs are exponential, they're gonna go out, they are gonna keep on looking for something on you until they find something. So again, you should be thinking what is out there from me, what is generating my profile in that world, you know, what is a data point about me, and what inference can be drawn from it? Because that will all be your digital footprint in extremis. Because the LLM will hoover it all up and it'll make a decision. Almost the worst case scenario is if there's very little, it'll probably draw a bad decision. And it won't be on your terms, and you can't proactively kind of help nudge that decision. So it's it's a good business case for building profile, doing things like podcasts and being out there in a way I think makes a a lot most of us uncomfortable. But it's taking your own destiny into your own hands to say this isn't the digital me.

SPEAKER_02 17:56

Yeah, no, I I personally completely agree. And um, you know, I don't know what your your sort of uh persona is like personally, but for me, um I I really enjoy doing this podcast, it's one

You Versus Your Digital You

SPEAKER_02 18:07

of the best things I do now, but it's not natural for me to be sort of quite extroverted in that sense, and actually, podcasting is quite a good medium for that because I can look in your eye and have a conversation and we can get lots of material um from that, so it kind of kills two birds with one stone. Um, but yeah, obviously not everyone's comfortable with that kind of um approach, is it? And and particularly obviously it's a generational thing, I think younger generations are more familiar with um having that digital persona and talking to camera and and and and seeing it as an extension of themselves almost. Um so so yeah, how how do you how do you think about that sort of transition if we call it that? I agree with you on where we're going. I think that's inevitable, but that um that the transitions are never linear, are they? They're always sort of slightly blurred.

SPEAKER_01 18:51

No, they're they're not linear, as you say, but the the the issue I think people need to think through is if we follow the rationale that automation, if you want to call it that, so you know, bigger use of data to and then be automated through LLMs or whatever comes next, that means that the humanity of what we do becomes amplified, and the humanity therefore is founded on relationships and other people's opinion and the network you operate in. Now, business always has been about your network to a large degree. Most roles are, whether you're hired for it, whether you operate within it, are network-dependent. So you have to be out there. So again, get to our rationale. If the administrative side of your role is being taken care of by some form of bot, therefore you do need to be out there more to do more of that type of role. Crudely, I didn't articulate it particularly well, but you see my point. I think we need to be more aware that the human element will require networking, relationships, profile building. That's what a lot of people see as being kind of sales issue and marketing will empower it. And I think most people would agree once you're senior in a business, sales becomes one of your full-time jobs and responsibilities, regardless. Yeah. Right? And for the younger people who have been happy, you know, happy to work side by side with AI, they should always be thinking about their networks.

Networks, Agency, And Senior Roles

SPEAKER_01 20:11

Yeah. And what that will get them in the next maybe not 10 years, but 15, 20 years, your network on the whole is always going to be everything, isn't it?

SPEAKER_02 20:17

Yeah, yeah, no, absolutely. And um, and to that point, you know, in the world that we're in where AI is exponential, the only thing that we can do is be human. And actually, in that we should lean into our humanity. And you know, that's what we we try to do as business, you know, try to engage with um human conversations and it and emotions and the messy stuff which is human, which isn't isn't data. And I think in that as well, there's going to be a big thing about proof of identity, isn't there? So and that's where this digital persona can really help people. Um digital clones and and avatars and other things that that may look to imitate people's. Which is terrifying.

SPEAKER_01 20:55

So I I also have a conversation with um the guys at EGU about this, and saying that maybe the only way to prove that it's not a bot is if you're in the meeting room. So suddenly face-to-face meetings become not there's a high premium attached to them as validity or validation of that you are that person who says you are with that criteria and that experience. But to your the point you raise around data points and human interaction, so that's at the heart of what I'm trying to do in my business, is actually I think that sales, that process of networking and discussion, does become data points. So we're hoovering up and harvesting data via emails, phone calls, PowerPoints, and so on, because they're all the largest part of the sales journey. Yeah, right. And sales journey is usually trackable to a certain degree, but when you pour it into that world of interaction, you lose track of it and you hope it surfaces in a CRM, right? Well, I think now, what I now know for the power of the AI and data application is that messy world of meetings and emails and back and forths becomes trackable. And understandable. And once you've had a degree of human intervention to explain what good is in that context, and more importantly, what bad is, the AI can make at scale dive into those different

Identity, Clones, And In‑Person Proof

SPEAKER_01 22:12

data points and start to give you a view to say, well, actually, of this many conversations, X number are useful. X number could be optimized. And it's taking a big data view, hopefully, of all those seemingly uh innocent or you know happen-stance conversations that could be holding an opportunity to cross-sell. Or they could have been opportunities to express a proposition or you know, look for your next role. We want to try and surface that so businesses could say, actually, we can become more productive and increase our sales rate, our sales conversion, and there would be such power in that, even if you're thinking about incremental increases. Yeah.

SPEAKER_02 22:48

Yeah, totally. So, so stepping back from this for a second, then, I we I think there's a lot of alignment between us on this, and I'm sure some people out there will be nodding along, and others maybe who aren't as far potentially along this journey will be sort of wondering about some of this. Um, but what what do you think, what do you think bringing this back to the physical world, back to cities, um, what do you think this is going to have as an effect on cities in terms of how people interact, uh, how people work, how people live?

SPEAKER_01 23:19

Well, um, I think the there's an example I gave at Place Camp. I think when you apply the rationale that we're we're discussing, this idea of you know everything's a data point in theory, structured and unstructured data. What becomes very interesting, I believe, is the way places that can then be tied to data points and provide information to the user, but it takes one or two conceptual leaps, and they're not they're not massive. So, one, the way we interact with a space, whether it will be through an earpiece, our phone, your glasses, you'll have an AI agent with you as you experience a place. And if that agent can draw data from the area around you, which it will, because the agent will, you'll enter on a base, a geo basis, enter an area, data, different data points for that geo area will start discussing things with you. So say you go to um culture

Surfacing Sales From Messy Interactions

SPEAKER_01 24:09

in a yard or whatever or whatever. Wait a minute, well, you come go to the strand. Suddenly the shops there start proposing deals and information to your agent as you're walking through that environment in real time saying you're interested in this, you're interested in that. Your virtual agent will know that you're shopping for this, or you're interested in this, or you're like these brands, it will start making decisions for you and helping you. Now, that sounds like it rather removes the agency from us as humans, but it's to enhance the experience, not to remove it entirely. So there'll still be the moment of serendipity that you want to walk into a brand or a shop. But you know, you've got the venerable RAC here. If the RAC says, right, there's a coffee morning or a networking morning on for real estate development, you as an individual should have a flashing light on your agent going, right, you should stop off here for a coffee today. You're in the area, this is an open opportunity. That's a very exciting potential for placemaking, but it's really dependent on different sets of data from both sides meeting. So you as a user need to have instructed an agent as to what your um interests and your preferences are. The RAC needs to have provided the data to be accessible in an interface, right? And said, We're doing these, we're doing these events, there are these promotions, we want to have this type of user. If those two things align, that's a great um additive experience to the whole placemaking journey that we have. And it will take a while, and I think some of it will be a little bit unerring for people because that idea of sort of you know control is is is quite is clearly a big one. But what it also, therefore, again, but that theory, what it removes is the need for the web and websites to be an interface because your agent will be directly interfacing with the data from the RAC, it will call up information about that event or that shop or that promotion. You only need to look that up. So the intermediary of a Google or a search, that's done directly through your agent interface. That's an exciting world for placemaking and developers.

Cities With Agents: Place As Interface

SPEAKER_01 26:03

You therefore think about the environment you're creating and all the different data points that represents. Who do you need to show those pieces of data to? Are they intelligible and understandable? How do we do that and structure it? Because when they're talking to one another, it's going to be really cool. Uh, it's gonna take work from both sides to make that happen.

SPEAKER_02 26:21

Yeah, yeah, fascinating. Um, so to put you on the spot a little bit, you've kind of talked about the impact on cities, streets, and probably retail there. If I was to take the other big food groups within real estate, so housing, workspace, industrial, let's say, maybe just take those in turn. What what would be your gut feel on the future of those, how how they may be affected by everything we're talking about?

SPEAKER_01 26:47

That's not um that's not an inconsequential question.

SPEAKER_02 26:50

It's not, no.

SPEAKER_01 26:51

If I take so a piece of work that we've been doing, which I found fascinating, is we looked at a commercial REIT and we looked at the various elements of disparate data within their universe. And just off the top of my head, within that, there are alarm data, uh, AMPR, car parking information, footfall, security, uh, then desk occupancy, unit, storage units, and so on and so forth. They're all different data sources, all sitting on different servers, all in legacy software. Our role is to try and roll that up into a surfaceable UI. So actually you could start to see and interrogate those different data sets and understand them. What's my point in raising that? In everything you mentioned from housing to sheds, it's thinking through what do we have at our disposal to help us help our customer, whether it's investor on one side, you know, homeowner on the other. For developers like a Barkley Homes or anyone in that world, what are we looking at in that environment around us? Not only to help us during the building process, but then post-occupancy, what can we do to understand the way these spaces are being used? How do we optimize that for future developments? How does that inform future

Housing, Workspace, Sheds: Data Everywhere

SPEAKER_01 28:05

sales campaigns and marketing campaigns? That that is almost limitless in terms of what you could pull from open source now, in terms of footfall, car traffic, uh, local planning, you see what I mean? Any of those, any of the above. But when people start to cotton onto this ability, I think, wow, we could overlay so much there. What else is out there for us? Businesses that would never have considered themselves data businesses will become really hungry for different sources like that. What what what's an additive to our business to layer in? And sheds being a simplistic example, but very complex at the same time. They're about swiftness of adoption, ease of use, ease of transformation, but all of those can be captured through data, probably now, to enhance the overall experience and therefore speed to market and yield and so on.

SPEAKER_02 28:49

Yeah, yeah, that's super interesting. And um, within this, what we're talking about is uh an increase in in data and uh and actually an increase in value within that, aren't we? And I think a lot of people listening to this would be really interested in well, how is that value captured? How is that value distributed? How if you're a if you're an asset owner, how you know, how do you get a share of that pie? And also if you're a citizen within the city, how do you attribute value to that? How how does value not just all accrue to the centralized technology company? I think this is a really, really interesting aspect of of um of value. And and you know, if you take it slightly bigger than that, you know, if we're in a world where um, you know, whether it's universal basic income or whether it's sort of of a big shift in in and where value lies within the economy, away from knowledge work, service work, into um some of these, at the moment, probably quite hard to grasp, areas of value. I think that that whole that whole topic is really quite interesting. I don't know if you have a view on that at all.

SPEAKER_01 29:57

Well, that's not a small question here, because the you you Yeah, because you're going through quite several theoretical leaps, which are great, but I I'll try and go

Who Captures The Value

SPEAKER_01 30:05

back a couple of steps. So the first point, which I think is really interesting, is who wins from all this? Now, to to answer that question, I think there's the idea that well, firstly, some businesses are erroneously trying to build an LLM or teach an LLM. Like you're not gonna outcompete ChatGPT or Google, there's no point, right? Because they're training off multiple, multiple, multiple data sets, and therefore there's accuracy and veracity within all that. However, your approach to AI will be highly dependent on what your business is represented through data, right? Data in its broadest sense. And by that, what I mean is the way we work, my business works, is we want to give people the best access to their best data and to set an AI within that data world, within those parameters. Because when it's operating within your parameters and your function, you will get an accurate, up-to-date, honest answer about what you the problems you're trying to solve. If you're in the open AI environment, it's going to draw in all other sorts of information and noise, and you'll get hallucinations in accuracy and so on. So, what we've learned is you've got to narrow this down and create parameters or a box around it, in short, to get accuracy. And therefore, you need to keep topping that box up with more information and so on, specified to your business. So, your point is who wins there? AI, the AI itself, in that world is a commodity because it's just coming into your business to enhance your business, right? So you are winning as long as you're a subject expert in your area of you know your area of knowledge and you're bringing more knowledge into that business all the time. Google can go and win over there with its Gemini and its open AI. You will win from bringing that AI into your business and applying it. What you don't want to do is be necessarily informing them by giving all your proprietary information up into that open environment. Because you're just training their models. You want to retain it within your world. So there's risk and opportunity there. But I think therefore the businesses who are open to this and are keen

Private AI Boxes, Not Public Models

SPEAKER_01 32:04

to progress, especially early adopters, they will take huge wins in efficiency. Not because they'll be getting all their answers from copilot, because that's clearly not going to work, and most businesses are struggling with that. It's for those guys who structure their business in a way that can be understandable to AI in a private, risk-free way. Yeah? Uh in terms of the the bigger leaps there for to society, which is a is a very big topic, I wouldn't like to presuppose because I like you, you know, I've watched Terminator too many times. It feels like there's too much risk around a lot of this. However, the AI that we're all you know aware of now is progressive and it is based on a version of our truth. It's not learning per se. It is not capable of independent thought per se yet. And maybe San Almond's got that type of sleeve, I'm sure that's what they're heading towards. For us, it is a sort of a tool to sort through our existing data and enhance it and try and make it better if we guide it in the right way. That for me doesn't pose a great risk, but a huge opportunity. It's somewhere off that idea of a universal basic income where suddenly we're all freed up to have loads of leisure time. I think it will ultimately create more interesting work. And I feel quite optimistic. But I do feel there there will definitely be some short-term paying as industries recalibrate around that. And that because of that, you know, some administrative jobs go, research jobs, entry-level jobs go, some senior-level jobs go because you just don't need as many people to do the same thing they did before. But you will need that experience, you will need to enhance data and you want to be creative, so there's opportunity there as well.

SPEAKER_02 33:47

Yeah, yeah, yeah. No, totally. And um I think on the on the value side, again, uh a lot of this is is speculation. And um, I I think there is an opportunity for um yeah, if if you think about all of the data sources that will come from people being in buildings, for example, that that the the owner of that

Built Environment’s Data Reality Check

SPEAKER_02 34:05

building is instead of just charging rent, could actually kind of draw some sources of value from that maybe over time that they can supplement the rent or even they can drop the rent because the value is actually in all of these other data sources. And then on on the individual level, uh, in the same way that you know, if if um you know, if something's free, you're the product is the same, isn't it? But you know, if if if we're going into a world where you know your your biological data on your Fitbit or the the way that you experience a space or the way that you shop and and as you mentioned, you know, your your AI is is integrating with the world, there is a two-way value proposition there. One is that you're getting value from the AI, and the other is that you as an individual are giving value to the AI, and and how can that actually accrue to some form of whether it's financial value or whether it's coupons or tokens or you know, whatever it might be. I think that that is a really empowering version actually of this is that the AI is is is in symbiosis with us, and actually the more the better the AI gets at drawing more sources of data, the more that humans and owners and other people can contribute that data and and seek some form of reward from it.

SPEAKER_01 35:15

Yeah, absolutely. But then in in the built environment, putting a question to you, are we good enough? Whether it's you know, you mentioned rezi to commercial, built-to-rents, sheds, are we good enough at really having the data points to understand why our customers do what they do? Well, regardless of the type of customer, it could be an investor, could be an end user, could be a company buying to a shed or taking a lease. Why are they doing that? How does it perform for them when they get it? Now, I I haven't seen enough evidence. I know there's people out there trying to solve this, but is there enough data collation points and then enough integrity and honesty about that to really inform decisions? I would challenge that.

SPEAKER_02 35:55

Oh, totally. I don't think we're anywhere near what we're talking about. Yeah, yeah.

SPEAKER_01 35:59

And when you compare it, it's clearly a lot easier in the SaaS world or whatever those businesses, um, technology businesses are selling, they get a lot more because their the data points are accessible to them. Because the way we interact and download SaaS and then play with it, they're all trackable, like using a website. In this world, you know, that will become increasingly trackable. But the industry still hasn't got anywhere near solving that.

Smart Buildings Or Smart Agents

SPEAKER_01 36:21

Internet of Things was supposed to be something that moved towards that. Turns out the data points are hard to collate, difficult to implement, expensive, and then ultimately it's not being used very well, right? Even though the potential is there. So we have hurdles to cross before that becomes meaningful. But if the rate of change is such in the other sectors and other parts of society, surely built environment will have to accelerate to keep up. Purely because everyone else will be making such database decision making, whether it's through the agent or whatever it is, that built environment will go quite, we we've got to stop and solve for this. Yeah. Yeah. And that's that's the exciting point of some nascent discussions I'm having with real estate businesses, is to start trying to solve that. Yeah.

SPEAKER_02 37:01

Yeah, no, I think that's really, really interesting. And also, but there is a you know, a counter to that is always that you know, the built environment is is often just the the backdrop. And actually, learnings from technology evolution in the past is that actually trying to integrate fast-moving technology into slow-moving assets isn't always the right approach. The very basic example is the the plug soppet that had the USB A in, and now everyone's got a USB-C. And that's a very minute example. But I think there's something in that, isn't there, to the extent that how much do the physical world actually need to become smart versus it just being the backdrop to very smart things that happen, you know, within it.

SPEAKER_01 37:42

Yeah, almost regardless. But then the example I would use, because I've always been mildly cynical, is the the building website. Right? So nearly every big uh building has now a website or some kind of digital interactive platform, and and quite a lot of BTRs now have those two things as well, right? But the

Wrapping Up: Practical Next Moves

SPEAKER_01 38:00

actual user uh count is small. The actual ways that you you harvest tasks through that or the way you would glean information is going to be poor, therefore, as a result, because it's not additive to the whole experience. You're not you're not deriving much from it as a user. You may have to log in or get a QR code, maybe get a free lunch or something. But I I haven't seen any data to suggest that people are really using them and they really get utility from them, and that the person who's provided the website gets much utility or data from them. So that's the kind of journey the sector will have to be on if it's going to match this challenge of us all using virtual agents and therefore wanting to interrogate a building experience through a virtual agent. You know, that the building like this one here will have to provide data points on occupancy, on um easier accessibility, you know, with the traffic bad here today, all those different things. My agent will be trying to solve that. And therefore it'll be reliant on the building and the building owner, therefore, providing the data. Yeah. You know, and that will become, I think that will become a choice to say, well, actually, I didn't know if there, if there was a deskway meeting available here, I'm not going to sign up today. I'll go to that other building over there. And then that will become more of a consumer decision in due course.

SPEAKER_02 39:07

Yeah. Super, super interesting. And I could literally carry on down this rabbit hole with you all day. But pulling out and sort of wrapping up then, a lot of I think what this leads to is um, you know, is a value-space decision for individuals, companies, and also if you want to talk about it precisely about kind of you know, what direction do we want to kind of take this in? And I think we do have a lot of agency within that, and I think I'm very optimistic about the future. Um, but if you were just maybe to like sort of leave the listeners with something around that about kind of the actions maybe that we can take day to day um on a personal level to to really get the benefit from this, but b kind of shape it in a way that we think is positive for the future.

SPEAKER_01 39:53

Well, um that's a question I should have prepped for. Um I think the so a lot of the the kind of status quo advice around an AI-enabled future is think about repetitive tasks that you have and that can be easily automated therefore because they're in essence robotic and process driven, and you can look at you can map them on a workflow. And I would I wouldn't challenge that because it's true, and you can do that, and that's a that's a simple use case for now. But I'd actually go a step further and say, well, think about the things which aren't robotic, which are creative and exciting but difficult, therefore. What are you gonna pull into that environment, whether it's your own GPT on the uh when you log in, whether it's using Claude, but what are all the things around you that you can use to build up a portfolio or a picture of you, your role, how you excel in that role, but how you're gonna make it better? Because suddenly you will you will you have it now, but it's only gonna become increasingly so. You'll have the power to enhance what you do through multiple different sources of information. And with information being fully democratized, retaining knowledge won't be the key to professional success. It'll be much more about your ability to synthesize and see through if I can pull this in and this and this and pull these strands together, I'll have the mechanical machine-based uh help to pull this into a shape and I'll apply my logic and rationale and experience. That's a pretty exciting place, right? So that that undermines everything we would have learned back at school and university about retaining knowledge, and it becomes much more about expression of thought and thinking. And for anyone who leans into that from a lateral thought, creative thinking point of view, that's really exciting, right? It removes administrative burden and it starts to overlay creativity on almost every job. That's that's got to be an exciting future. What I think is fun for us is how can built environment double down on this now and enjoy it and get ahead of it because it's a really exciting time.

SPEAKER_02 41:50

John, thank you very much for your time today. As I said, I could talk to you about this all day, and I'm sure we'll love to have a round two sometime because this is it's been a fascinating conversation.

SPEAKER_01 41:58

Thank you very much for having me on. Cheers.

SPEAKER_00 42:01

Thank you for listening to the Grow Places podcast. For more information, visit growplaces.com and follow us at We Grow Places across all social channels. See you next time.