The surprising ways artificial intelligence is reshaping real estate demand

While reading the headlines might lead one to conclude that AI is the death knell for office buildings, JLL research shows the reality is more nuanced. In fact, in some markets, it’s entirely the opposite.

Rather than retreating from physical space, tech companies building AI infrastructure are expanding aggressively. In Sydney alone, the technology sector's office footprint has nearly doubled over the past five years to over 400,000 square metres. In markets like San Francisco, AI firms now account for nearly 30% of all office leasing.

But there’s a paradox: while AI creators gobble up premium space, traditional businesses are using the same technology to automate back-office roles and potentially ‘do more with less’. So, what's the net impact on central business districts?

"I think the public narrative really conflates how AI is going to impact labour markets versus its real estate impact," says Ronak Bhimjiani, real estate economist for Australia and New Zealand at JLL. "They're not actually tracking on the same trajectory."

In this episode of the JLL Perspectives podcast, host Rebecca Kent explores this tension with three experts navigating AI's real-world impact on commercial property. The panel examines which industries and job roles are most exposed to automation, what the ‘co-pilot economy’ means for workplace design, and why investors need to think in scenarios rather than single forecasts.

The conversation also turns to what this divergence means for property owners. Research from Yuehan Wang, who leads JLL's Global Technology Research Programme, identifies a widening gap between premium ‘experience quotient’ (EQ) assets that are thriving, and older B-grade buildings falling into a ‘disruption zone’.

With supply pipelines historically low across major Australian markets, the implications for vacancy rates, rents, and capital values are profound.

"All landlords are concerned about this," admits James Montague, head of Office Leasing in Australia. "But what we take comfort in is this is an evolution of workplace, no different to previous recalibrations. The net effect is a growth factor rather than a detractor."

The discussion also tackles unexpected questions – from the shocking cost of AI tokens to whether deepfake technology will make face-to-face meetings the only trusted form of communication by 2030.

Listen to the full episode to hear how landlords, occupiers, and investors should be positioning themselves for an AI-driven future that's already reshaping our cities.

Featuring Yuehan Wang, global technology research programme Lead; Ronak Bhimjiani, Real Estate Economist for Australia and New Zealand; and James Montague, Head of Office Leasing in Australia.

Perspectives podcast: AI was supposed to kill the office. Instead, it's sparking a space race

Rebecca Kent

welcome to JLL's Perspectives Podcast. I'm Rebecca Kent, content director at JLL, and joining me now are Yuehan Wang, who's leading the Global Technology Research Programme for JLL; And we have the real estate economist for Australia and New Zealand, Ronak Bhimjiani, hello, and head of office Leasing in Australia, James Montague.

Ronak, I'll start with you. Someone pulls you up at a dinner party and says, AI is the, is AI the end of offices or what? What's your short sort of simple response to that?

Ronak Bhimjiani

The one thing I like to kind of point out very quickly, if somebody did answer that question is, I think the public narrative really sort of conflates how AI is going to impact labour markets versus its real estate impact. And that's what we're involved in on a day-to-day basis. They're not actually tracking on the same trajectory or the same curve.

The early call that we're seeing based on the data that we're observing is that so far, at least in the markets that we're tracking here in Australia and even globally, which Yuehan will talk to, there's been a massive upside story for broader take up among key occupiers, tech occupiers across office markets.

Yuehan Wang

(Since) back in February this year, this AI scare trade is really a thing. Like people are afraid. People are genuinely seeing the capability of Claude, of ChatGPT and all the other things that's not just involving a chat bot – you say something, it gives you something back.

But the whole, let's say, office collapse scare trade is based on the, let's say, AI agent. When people discover that AI not only can encourage some tasks back to you, but it can go out and do things, execute those tasks, which makes them much more powerful than people have expected a few years ago. And so that means genuinely human exclusive tasks will keep shrinking. And this stable relationship between head count and work output, they start to decouple, these two.

And so the thing that people, the methodology that people are used to saying that, okay, if this industry grows, its demand for space will grow proportionally. That is being challenged. And I think that's the motivation behind all these fears. However, as Ronak rightfully points out, that does not mean it directly shrinks headcount as a fact. So there's the theoretical capability of exactly what people are able to achieve with using AI. These are two different things. And then what happens with the new way of working and space demand is a different thing, and space demand and realistic performance is yet another set of relationship.

Rebecca Kent

Okay, so let's dive into some of the industries, perhaps where we see the impacts of AI, in terms of maybe headcount. Yuehan, is that something that you can comment on?

Yuehan Wang

Yeah, so I say that I'm glad that you raised this question because looking at the labour force changes, you can do that from a few different lenses. You can do it on a city's level, you can do it on a real estate portfolio level, which is more charted by the location or the space.

But I would say that if a market is, whether a market is going through expansion, opportunity focus, or is it really that, let's say, high turnover, just.

not overall shrinkage of demand, but just changing entity of where the demand is being driven from, that is really very much influenced by what industry is present or dominant in those market and what kind of employment structure is currently in place for those markets or sub-markets.

Rebecca Kent

James, what are you seeing on the ground? I assume the impact or maybe the conversation or the types of or level of conversations you're having are different with our landlord clients in a CBD versus those who have, you know, more exposure to maybe fringe or other types of markets or even geographically across the country.

James Montague

Yeah, I don't think that any investor landlord, irrespective of geography or where their asset is located, is not worried about what AI is going to do to the occupancy of their buildings. I think if we were to ask in every single research presentation,

that we do, how many times is AI, what effects are you seeing AI have? If that wasn't asked in every research presentation, I'd be very surprised. So I think all landlords are concerned about this.

Where we take some comfort is this is an evolution of workplace, no different to a recalibration of

being in an office to being in open plan. And we've gone through work from home, work from the office, and the evolution of workplace very recently. So it's a compounding effect of that uncertainty. And there's no question, headlines that read

with a negative story get the most attention. No one's getting excited about the fact that headcount has remained stable because of the adoption of AI. But people want to hear about the things that have happened where the big needle moving changes.

And talking to the state heads across the country, unanimously, there's no broad brush impact of headcount from AI being a disruptor to workplace. What we are seeing, and maybe this might be too broad a statement or a bit contrarian, but

I think that a lot of businesses are taking the opportunity to perhaps tighten up their workforce and using AI as the scapegoat in order to be able to get the exact working population of their business how they want it and make it more efficient through the use of AI.

What is interesting, we've seen a lot of headlines about AI being implemented, and then a shock when they realise how much it's costing versus traditional employment. So that recalibration is not only a workplace and headcount one, but a financial one.

We're at the start of this change. I think from an Occupy standpoint, and Ronak pointed out that there's been significant growth, particularly in Sydney.

There has been tech businesses that have shrunk and there have been tech businesses that have grown. Professional services are probably more stabilized, but maybe have less of a growth mindset than they do, than they did have pre-AI coming into the conversation. So there may be more

cautious about what roles they need and don't need. But if on balance, if you looked at the net absorption factor or how many offices have been populated by technology businesses today versus five years ago, there's a huge wave of influx of people that are now occupying our cities that weren't.

So the net effect of that is, it's a growth factor rather than a detractor.

Rebecca Kent

Ronak, do you want to jump in there?

Ronak Bhimjiani

Yeah, I think just the one point I'd like to add on that, and James, you've hit the mark there in terms of, you know, we have seen this broader uplift in take up among key technology related tenants across key markets. You know, gone are the days I feel where you can expect an upturn in a cycle, an economic upturn in cycle or a job market boom to essentially lift all the boats on the tide. Like the secondary grade market is a good example where historically, when we saw an upturn in economic performance and you saw more employment, even the secondary grade asset that starts started performing quite well. Moving forward, that will become much more of a challenge for those B, C and D grade assets to naturally perform well, whilst your premium and A-grade assets will continue to sort of out-deliver, and that gap will essentially widen.

Yuehan Wang

Yeah, and I want to just add on to that, because one thing that's the driver behind some of these is not only which industry is present, it's about which corporate function, like which part of this industry is located at certain places. You would say that a call centre or a customer service department might not necessarily sit right next to their C-suite and their strategy department. And so that is, again, a next part that drives that divergence. Because we have to admit that not to say like that we're purposefully downplaying the impact. No, that's not true. There are certain drop rules that are more disruptive than others, namely, let's say back office roles or these customer service roles are on the frontier of being, let's say, like automated.

But they are like, this is where it really becomes a real estate question is where do they locate? And what does that mean? That's like that connects her into the space demand story.

James Montague

I don't think that that is a huge divergence from the trend that already existed. So like we were and the topic of conversation, clearly we're focusing on technology, but all across industry where tenants are looking to attract and retain the best talent,

they're moving into better quality buildings. So it's not unique to the technology industry. And I agree exactly with your comment. If you have a function that doesn't require high cost real estate, then you're not going to put them in the best building in a capital city. They're more likely going to occupy cheaper real estate further away because

that function doesn't need to have the connexion into the corporate occupiers in the CBDs

Ronak Bhimjiani

Yeah. So if you think about the way AI is going to impact what we do on a day-to-day basis, I think there's going to be much more of a spotlight on client-facing type roles, where we can, you know, essentially use our EQ to our advantage. And the second part of that is data becomes key.

and the infrastructure or the hardware behind it. So if you're looking at buildings or office assets more broadly, I think tenants are going to be looking for those assets that deliver a frictionless experience, which they're doing that already with a flight to quality that we've been talking about for a long time. But this adds another layer on that. data and infrastructure, and that whole collaboration story around EQ-related roles and client-facing roles. So it's just another layer.

James Montague

Can I push this a bit further out and be a bit futuristic in this question? Where you have deep fake technology that's getting to the point where people can operate an AI client of themselves in a Teams meeting,

Do you get to a world where face-to-face meetings are the only way that you have an ability to communicate where you can trust the output on the other side?

That might be a little bit too far in the future to predict, but that, like, I'm picking up on your point about EQ, and you know, we're now having phone conversations with AI, which, like, I think you can mostly still tell, but it's getting to the point where it's harder to tell and you are seeing deep fake images come out that are so lifelike, they're hard not to believe. So at some point it gets to a sophistication, a level of real person. The only way to remove that doubt is a face-to-face interaction.

Rebecca Kent

And, therefore …?

James Montague

The knowledge-based economy that we're describing and the high value outputs of the people that are employed by businesses will need to interact together. So it will only perpetuate the need for CBD occupancy and continue to drive that need for an office footprint.

Rebecca Kent

If I was to say, well, you would say that, James, because you're a leasing agent and you have a vested interest.

James Montague

Absolutely right. Yes.

Rebecca Kent

We would be absolutely right. Well, do we know enough? You know, Yuehan, in your research, do we know enough about what the future looks like in this space?

Yuehan Wang

Honestly, there are … this is the interesting bit, like experts or world renowned professors in doing labour economics or technological change studies, they do not have an agreement at an aggregated level, where we will land.

There is like part of this, let's say the community is more thinking this just as all the previous, like major technological breakthroughs will eventually bring more jobs with like a better economy. There are others who are afraid that the turbulence to achieve that in the middle will break. things. They're all valid, but this just shows how difficult it is to like discuss this definitive aggregated results for everyone globally. But what I would say is that this type of discussion is.

less relevant if you're operating real estate locally or making any real estate decisions locally because like the story as we find here is it is not really universally like creating impacts like evenly for everyone. The divergence is really the story.

And for anyone who actually own an asset or operate in a specific market, my recommendation is that while the future remains unknown in the aggregator level, it's important to understand what's happening in a specific market and that you have much more reliable data

that you can look at. Just as we discussed, you look at the industries, you look at the job role and employment structure in that, you look at the investments that's currently going into it, expecting some form of expansion or growth need coming from that. And I would also like to point out

like another side of the story. Everything we discussed are the demand side story. We also have the supply side story that we need to add into the mix because that would either amplify or mitigate this effect on demand.

I will leave to the other gentlemen to explain like what's happening in the Australia and New Zealand market, but globally, let's say in major metropolitan, especially in like East Europe or the US, under supply of quality building,

is an issue. And it's something that's getting more and more serious. And overall, the age, the ageing of the premium office stock is another reality that we need to grapple with. So while maybe even there is some softening of demand, there's even

less supply that's available, which means like we're all pretty familiar with that formula. It means you can even expect higher rent and like pretty competitive like bidding in those markets.

Rebecca Kent

So it cushions the impact, if any, and wherever it is. Is it that simple, Ronak? Do you think in markets where there is limited pipeline coming through, then we won't see the, that dramatic, that scare narrative playing out.

Ronak Bhimjiani

Yeah. So, in Yuehan's research, the research that we've done here in Australia, we've taken three or four scenarios around the implications of AI advancement on the workforce. And we've actually formed our analysis based on four different outcomes as opposed to one. So we can go to an investor and say we don't know what the answer is, but you can take your pick based on what you think the trajectory of AI is going to look like. Do we think it's going to be an augmentation tool? Do we think it's going to be a replacement tool? There's loads of outcomes that we can be discussing here.

Now, the supply side of the equation, Rebecca, is very very much prevalent and very interesting on an Australian landscape because every market that you assess in Australia, whether it's Sydney, Melbourne, Brisbane, Perth, Adelaide, and Canberra as well, which I want to talk to as well, there's a low supply pipeline story or narrative across the major CBD markets. The only outlier there is Canberra, where there is a bit of supply in the pipeline that will be coming.

However, we've also assessed these markets based on what we think the AI exposure score is for every single market. So Sydney and Melbourne, for instance, have a slightly higher AI exposure score in terms of AI's impact on the labour market, the white collar labour force, relative to markets like Perth, Adelaide, and Canberra. Right? So right there, you're going to see a broad range of outcomes under the co-pilot assumption.

But under all the outcomes, we broadly are seeing lower vacancy rates.

Yes, they might be slightly higher in terms of the forecasts relative to what we produced previously, but still a lower vacancy profile because we're still seeing a churning employment growth. And the supply pipeline across office markets in Australia is fairly low.

Rebecca Kent

And Yuehan, are there any specific markets you can call out globally where you've seen that sort of tension between supply and demand?

Yuehan Wang

Let's say New York, for example, and some major markets, San Francisco, and also in London. These are all namely places where they have a lot of AI leasing happening at the moment, and they are diverse enough.

And they are gateway enough that even though, let's say, there's internal bifurcation happening with financial services, with professional services, these market don't lose those tenants. These market will continue, absorb, become like the face of those industries. And they suffer from a drying up pipeline and also like ageing building stock and the overall limitation in how much quality space that's available in that market. So that's still, yeah, that's still available. And so as a result from the US office markets, we actually seen a record high in leasing, bidding, and also in like the premium price. Since I forgot the, if it's a since COVID or like one of the historic high point is what we're seeing right now in there. So that just shows how much.

in at least in gateway cities. And I think Ronak mentioned how different those markets are, which is very true globally too. But I'd say for these type of markets, it's really the supply side being a very important part of why there's that high Like bidding happening.

Rebecca Kent

It's important to remember that there's a lot of other stuff like disruptive forces going on in the market, right? Like in terms of the broader economy, tariffs in APAC.

James, you were talking about some of the people you talk to in the market using AI as a bit of a scapegoat when there are other pressures. Is this kind of what you were getting at?

James Montague

Yeah, absolutely. It's widely publicised that AI is a disruptor for the workforce, rather than admit that your performance isn't as strong as you might like it to be. You might take an opportunity to reduce some headcount in some areas of your business and the impact has been from AI.

Equally, from a real estate decision-making standpoint, there's been a handful, and I certainly wouldn't call it a trend, but a handful of businesses that have...

reduced in quality to take more flexible arrangements in lower quality assets because the pricing of all of our prime grade stock has gone through a massive recalibration. So they're now retreating out of what was probably a deal done during COVID where

They got a cheap deal, and now they're moving backwards in quality and talking about needing more flexibility because they're not sure what AI is going to do to their business. Again, it's an interesting narrative to hang on to, to give some sort of decision-making Credence to a downgrade, rather than just saying, "Look, we like the business is okay, we're not gonna sack anyone, but it's not quite where we want it to be. We can't stay in this super high quality product. We'd like to, we wanna move backwards in quality and just make sure that we're managing costs, so it's rather like...

AI is the reason, not appropriate OPEX management. So there's been a handful of, but I'd say that they're unique situations and they're privately owned SME companies doing things like that. We're not seeing moves in the global, even multinational professional services firms where they're downgrading in quality.

For the most part, that flight to quality remains consistent and chase for high quality product, better amenity, better experience, that description of frictionless environment where you can start operating your day in the lobby of the building, move through the lifts into your office space and maybe into some outdoor environment.

all fully connected into your premises, that workplace experience still forms a big part of decision-making criteria when tenants are assessing whether they stay or go.

Ronak Bhimjiani

Yes.

higher clip. And there is clear implications on labour markets more broadly. So take Australia, for example. The long term historical average a long time ago was trend employment growth at 1.7%, then 1.5%, and moving forward in around 1.3%.

to three currently down to 65% and based on the Oxford Economics numbers that will be down to 63%. So there's a natural slow slowdown in employment growth with or without these external factors and AI at play as well. So it's important to distinguish the signal from the noise and let's not make that blanket statement by saying it's just AI that's infiltrating the labour market in a negative manner.

Rebecca Kent

Yeah, so unfair just to blame AI. James, when some of your clients say that you've sat in presentations and when they said, what's the impact on AI, what do you think they're really asking you?

James Montague

Will tenants shrink in size in our buildings? It's only that concern. I don't think anyone's worried about them evolving into a different company. What their concern is, is exactly what Ronak said. It's trying to find the signal from the noise because

a lot of the press articles are either about building new data centres and expansion of those mega tech companies or how professional services have laid off people because of AI. So it's kind of binary in the way it's reported.

There's

There are concerns about it, but when all of those landlords survey their own tenant customers, overwhelmingly the feedback is that we're not planning on having severe reductions in headcount because of AI adoption. And I think most Australian businesses look at AI as a way to do more with the same rather than do more with less.

And if you consider, and I look at our own business as an example, and we're a professional services firm, if each of our employees could be generating an additional 15 or 20% using just an extra tool that we give them, then every additional person we employ

has a compounding growth effect. So it serves as a growth enabler more than it does an ability to reduce cost because the other element we touched on it before is it's not cheap. So, you know, not every company can just roll out a full suite

of fully integrated AI tools. So we've got the luxury of scale and being a global business and we're giving out people access to it, but it's not this broad brush thing that everyone can just get a licence and start using it. So I think that adoption piece is going to be really interesting to watch because while these tech companies and

And I only scratched the surface, so I'll let Yuehan talk more with more certainty to this, but they've been on a capital raise environment, and now they're needing to generate revenue, and that's coming through charging for tokens. All the companies that have embedded and started using it have been shocked by the cost. So until it gets commoditised and it's cheap, which feels like it's a way away, then it's not going to have, not every single person in every company is going to have a clawed license. Certainly not in the foreseeable future from what I can gather.

Yuehan Wang

This is a great point. I think I'm lucky enough to sit in a programme where we look at not only these market shifts, but also how people use AI – the actual adoption of it. And it has been, I'd say, now a common understanding that generate expected results from AI is not easy.

And this year we were able to do another like AI focused survey on investors and occupiers alike. And the preliminary result that we start to see – this is all published content. It will come out later this year. But preliminary analysis already shows us that people want to achieve more of that revenue side of things, how we can help them generating alpha, generating more opportunity. But the results that they are seeing more now is very much on company operations and efficiency, just like doing things faster. It is not directly like allowing companies to achieve that expansion or that much of A cost save. There are multiple reasons in that that we won't have the time to dive into, but just overall, the biggest, let's say, barrier to achieving more results is that people gradually realise how important it is to have this data foundation, to have the correct context set for AI to do things, and also to customise those like agentic workflow that is very much bolted onto how your people operate, how your specific company operate, and all that takes.

investment and talent. And from the Future of Work Survey this year, I think one of the most striking results for myself is that for the very first time in the history of the 15 years of we run that research, skill gap becomes the biggest barrier for transformation, surpassing budget constraints. And it also just shows how much people don't feel they have enough capability in-house. This is the same for real estate investment community. And for investors from our latest survey, we see that

80% are hiring new people into their team to drive the AI use and another 80% there are overlap, but 80% are hiring an external partner to help them organise their AI adoption. So it's not that easy.

Ronak Bhimjiani

I'm hearing those numbers for the first time, Yuehan, so it's quite nice to see that that really points and supports that view around AI as an augmentation tool and used as a co-pilot as opposed to a replacement tool. And you know, the way I like to kind of see it is having these AI skills will be like having a driver's license.

Ronak Bhimjiani

You can't drive without a license. You can't work in some, you know, in some of the white collar roles without the licence to actually use AI in the right manner. So I think it's going to be a tool to help as opposed to a tool to replace across labour markets. Yes, there'll be a lot of musical chairs being being played around across the workforce. Work will change, some jobs will change, but as a whole or as a net basis, I think there was still that opportunity for growth and productivity lift because the key issue that we face here in Australia, and I think it's maybe a slightly different story for the US, is we've got a shocking productivity growth rate in Australia.

We've done nothing in the last 10 years to make our economy more productive. So I can clearly see a scenario where AI comes in, becomes a lot more embedded in our workforce, but the workers are still there to master it and use it as a tool to lift productivity.

Rebecca Kent

Amazing. Thank you. And Yuehan, I was going to ask you what we were going to see from you in a research capacity next in this space, and you've given us a bit of a preview. So that's great. Ronak, did anything surprise you when you were doing your research around AI and the impact on markets in Australia?

Ronak Bhimjiani

Rebecca Kent

James Montague

human interface and expertise locally is going to drive value and that I don't think will go away.

Ronak Bhimjiani

So what James is truly trying to say there is that Yuehan and myself would have our jobs pretty solid and safe there.

James Montague

All four of us, in fact.

Ronak Bhimjiani

Fingers crossed.

Yuehan Wang

I want to use that, piggyback on that, and do another preview of a research project that we have embarked on, which is trying to understand with all these ways, how does that mean a good office space design looks like? Like how can we actually design the space

robotics-fit, you need to think about that now when you design for that building to be built in that time frame. So just want to put that out there.

Rebecca Kent

Well, you'll need like a delivery door for drones, won't you?

Ideally.

Ronak Bhimjiani

A helipad.

James Montague

And just swap the EV charges in the basement to be robot charges. Is it as simple as that?

Yuehan Wang

robot chats or do they need their own elevators? Do they need their own like vertical transport systems? That's going to touch the core of the building now if we consider those factors.

Rebecca Kent

You mean we can't get in the lift with them and talk about the weather?

Ronak Bhimjiani

Yuehan Wang

Yeah, whatever we say are going to be off, that's for sure.

Rebecca Kent

You’re far too cautious, Ronak, disappointed in you. Well, great chat. We're going to end it there. Thank you so much. I think hopefully we've answered some of those landlord questions, James, that you're getting in presentations and you're going to have to repeat yourself over and over. This is brilliant. Thank you for your insights, Ronak, Yuehan and James.

Rebecca Kent

Catch you on the next episode, hopefully.

James Montague

Thanks, Rebecca.

Ronak Bhimjiani

Thank you. Thanks for your time.

Yuehan Wang

Thank you.