Discover how JLL built its own AI model trained on decades of data - and what it means for the future of property decision-making.

A Quiet Signal of What’s Coming

While most of the property world is still experimenting with AI, one of the biggest names in global real estate has already made its move.

JLL has over 60 AI initiatives, including its own generative AI model, trained specifically on commercial real estate data. It’s called JLL GPT, and it’s an example of a real estate company going beyond AI experiments and actually implementing it at scale.

The tech is built using decades of proprietary data (internal and external) to make smarter and faster decisions about property.

And that’s what most firms haven’t realised yet: they’re sitting on a mountain of data they don’t know how to use. Lease histories, occupancy reports, building performance metrics, fit-out benchmarks, tenant churn. It's all there - just waiting to be activated.

What JLL has done is simple but powerful: they’ve turned their institutional memory into an intelligent assistant.

Let’s take a look at what JLL GPT is exactly, the kind of data it uses, and the real problems it’s already helping solve.

What Makes 'JLL GPT' Different?

JLL GPT was developed in-house by JLL Technologies, the firm’s tech innovation division. It’s built on top of a large language model (LLM), but with a key difference: it’s been trained and fine-tuned on decades of proprietary JLL data, layered with public and commercial property datasets.

It’s being used across the company to assist with:

• Investment analysis & portfolio strategy for clients
• Site selection and location advisory
• Lease optimisation
• Sustainability reporting and ESG performance

The result is something far more targeted than a general-purpose AI tool. JLL GPT understands property language, market context, and operational nuance - because it’s been trained inside one of the biggest companies in the property industry.

So how does it work?

Here’s what makes it powerful:

1. the breadth and depth of data it uses - and;
2. How the AI turns that data into actionable insight.

Let’s break it down:

1. Property Listings & Transactions

📌 The data it uses:

• Commercial property listings for sale, lease, or investment
• Achieved sales prices, lease terms, time-on-market.

🤖 How JLL GPT uses it:

➜ Helps landlords and investors price property competitively

➜ Predicts time-to-let or sale based on market patterns

➜ Recommends pricing strategies using historical comps

🛠 Example questions it can answer:

“What’s the average lease term and rents for Grade A offices in Leeds?”

“Which streets in London demand the highest rent and have the shortest time on the market”

2. Market Trends & Economic Data

📌 The data it uses:

• Economic indicators: inflation, employment, interest rates
• Sector-specific development pipelines

🤖 How JLL GPT uses it:

➜ Forecasts market resilience or softening by region

➜ Helps developers decide when and where to build

➜ Advises on when to buy, hold, or divest

🛠 Example questions it can answer:

“How many new houses are due to be delivered in West Yorkshire in 2025.”

“How are rising interest rates affecting Grade A office yields?”

3. Tenant & Occupancy Data

📌 The data it uses:

• Space usage and underutilisation.
• Tenant renewal rates, break clauses, covenant strength.

🤖 How JLL GPT uses it:

➜ Identifies buildings at risk of vacancy

➜ Forecasts tenant retention

➜ Recommends space reconfiguration or repositioning

🛠 Example questions it can answer:

“What assets in this portfolio might need repositioning and why?”

“Estimate the vacancy rate of this building in 2027?”

4. Investment & Capital Markets Data

📌 The data it uses:

• Yield trends, cap rates, IRR by asset class
• Finance terms by lender type and deal size

🤖 How JLL GPT uses it:

➜ Matches buyers with opportunities based on recent deal trends

➜ Identifies where capital is flowing, and why

➜ Helps investors model risk-adjusted returns

🛠 Example questions it can answer:

“What’s the average yield on new-build office stock in West London?”

“Which type of assets are international buyers prioritising?”

5. Sustainability & ESG Data

📌 The data it uses:

• Energy usage, carbon output, EPC ratings
• ESG scoring frameworks and regulation timelines
• Sustainability incentives and retrofitting benchmarks

🤖 How JLL GPT uses it:

➜ Identifies assets at risk of future non-compliance

➜ Suggests actions to meet ESG standards

➜ Recommends green investment strategies with long-term upside

🛠 Example questions it can answer:

“Which buildings in Manchester risk failing 2030 EPC regulations?”

“What’s the ROI on a full retrofit to BREEAM Excellent?”

Why This Matters

What JLL has done here isn’t flashy.

It’s a strategic move to turn company knowledge into strategic foresight.

To build a system that understands their business, their market, and their clients better.

For the rest of the industry, it’s a wake-up call.

Most firms are sitting on goldmines of data they’ve never structured, analysed, or activated. And while the focus has been on public AI tools and productivity hacks, the real competitive edge will come from building intelligence on top of your own data.

JLL isn’t guessing anymore.

They’re analysing, modelling, and deciding - faster and with more confidence.

Final Thought

The next evolution of AI in property won’t come from chatbots, it’ll come from businesses who understand their own data well enough to train a system to think like them.

JLL just showed us what that looks like.

The question is: who’s next?