Skip to content
All resources

Blog

JLL GPT explained: what it is, who gets access, and why we think a specialist platform is the better option for CRE teams

Updated

JLL GPT is JLL's in-house AI assistant for commercial real estate. What it does, who can use it, what JLL Falcon changed, and why we think a specialist platform is the better option for CRE teams outside JLL.

Editorial cover for “JLL GPT explained: what it is, who gets access, and why we think a specialist platform is the better option for CRE teams”

Short answer

What is JLL GPT? JLL GPT is JLL's own generative AI assistant for commercial real estate. JLL announced it in August 2023 as a large language model built by and for CRE, and in October 2024 launched JLL Falcon as the AI platform that now powers JLL GPT and a growing set of JLL tools.

Is it the same as ChatGPT, and can firms outside JLL use it? No. JLL GPT is not OpenAI's consumer ChatGPT product. According to JLL's public information, it was built for in-house use by JLL's own professionals. We have not found an official JLL page offering JLL GPT to the public or to clients as a standalone product.

What should other CRE teams use instead? We think the better option is a specialist commercial real estate platform that works on your firm's own documents and systems, keeps sources visible, and does not sit inside another market participant. That is what Fifth Dimension is built to do. Book a demo.

What is JLL GPT?

JLL GPT (also searched as JLLGPT, JLL AI and JLL ChatGPT) is JLL's in-house generative AI assistant for commercial real estate. On 1 August 2023, JLL announced JLL GPT as "the world's first large language model built by and for the commercial real estate (CRE) industry", developed by JLL Technologies. JLL says it is trained on CRE-specific data from internal and external sources, has a conversational interface, and is positioned as a tool for human-led work rather than a replacement for professionals.

We first wrote about JLL GPT when it launched in 2023. This update keeps those arguments, adds JLL's later public information on JLL Falcon, and sets out why we still think a specialist platform is the better route for investment and asset management teams that are not inside JLL.

Source: JLL's JLL GPT guide (opens in a new tab).

Is JLL GPT the same as ChatGPT?

No. JLL GPT is JLL's own CRE assistant, not OpenAI's consumer product. If you searched for "JLL ChatGPT", "JLL chat GPT" or "chat GPT JLL", you are looking at JLL's internal tool, not a public ChatGPT workspace with a JLL skin.

JLL has said that businesses were hesitant to use tools such as ChatGPT "due to both security concerns and inaccurate outputs", and presented JLL GPT, trained on CRE data, as its answer. JLL does not say publicly which underlying foundation models, if any, sit behind JLL GPT or JLL Falcon, and we will not guess.

Can you use JLL GPT?

According to JLL's public information, reviewed in October 2026, JLL GPT was described at launch as a tool that would "initially be used in-house by our JLL global workforce". We have not found an official JLL page offering JLL GPT to the public or to clients as a standalone product. Clients can get AI through JLL's services and Falcon-powered tools, such as JLL Property Assistant and Azara Analytics, rather than by logging into JLL GPT themselves.

If you are an investment manager, REIT, lender or asset manager outside JLL, JLL GPT is not the assistant you can roll out across your deal team. The rest of this article is about what that in-house model of AI means for the industry, and what we think firms in that position should do instead.

What does JLL GPT do?

At launch in 2023, JLL described use cases including conversational space utilisation dashboards, AI-assisted client feedback and workplace planning strategies.

In October 2024, when JLL launched JLL Falcon, it said more than 47,000 JLL professionals had used JLL GPT, and that upgrades included image understanding, an enhanced knowledge base and 25 times more working memory.

In a June 2026 interview, JLL's CTO, Yao Morin, said more than 50,000 employees actively use JLL GPT, that employees have built 900 or more custom assistants, and that Falcon powers 40 or more applications. We attribute those figures to JLL's CTO rather than to our own measurement.

Sources: JLL's JLL GPT guide (opens in a new tab), JLL Falcon announcement (opens in a new tab) and the June 2026 CTO Magazine interview (opens in a new tab).

JLL Falcon: the AI platform behind JLL GPT

JLL Falcon launched on 29 October 2024. JLL describes it as its AI platform, covering multi-modal foundation models, data pipelines, security and privacy, natural language layers and analytics, and as the platform that powers the next iteration of JLL GPT.

JLL says Falcon is built on nearly two petabytes of its own transaction and operations data. Falcon-based tools offered to clients include Azara Analytics, Prism AI, Corrigo and AI leasing tools. JLL Property Assistant, an AI assistant for owners built on Falcon, launched on 28 May 2025. On 22 July 2026, JLL added Falcon-powered AI to its ATG Intelligence healthcare compliance platform.

These are JLL's public descriptions of its own platform. We do not have independent visibility into Falcon's models, training data or performance, and we do not claim otherwise.

Sources: JLL Falcon announcement (opens in a new tab), JLL's AI services page (opens in a new tab), JLL Property Assistant announcement (opens in a new tab) and the ATG Intelligence announcement (opens in a new tab).

JLL GPT timeline, 2023 to 2026

  • 31 July 2023: Fifth Dimension first published this article on JLL GPT.
  • 1 August 2023: JLL announced JLL GPT as a large language model built by and for CRE, for in-house use by its global workforce.
  • 29 October 2024: JLL launched JLL Falcon, the platform that now powers JLL GPT. JLL said more than 47,000 professionals had used JLL GPT.
  • 28 May 2025: JLL launched JLL Property Assistant, a Falcon-based assistant for owners.
  • June 2026: JLL's CTO said more than 50,000 employees actively use JLL GPT, with 900 or more custom assistants and 40 or more Falcon applications.
  • 22 July 2026: JLL added Falcon-powered AI to its ATG Intelligence platform.
  • 10 November 2026: This article was updated with JLL's later public information and with our current view of why specialist platforms are the better option for firms outside a single in-house build.

JLL reported 2025 revenue of $26.1 billion. Source: JLL's 2025 financial results (opens in a new tab).

What JLL GPT tells us about AI in commercial real estate

The launch of JLL GPT was a useful signal. It showed that large CRE firms see AI as a core tool, not a novelty, and that the data real estate actually runs on is mostly not sitting on the public internet. We still think that point is right. Horizontal, cross-industry products like ChatGPT are trained largely on public text. They are useful for drafting and general research, but they have not been trained on the offering memorandums, rent rolls, T-12s, lease negotiations and investment committee (IC) files that sit inside firms and data rooms. That is why we think vertical, specialist products are the better fit for real estate workflows.

The harder question is what kind of specialist product the industry should want.

We think there is a structural limit to the in-house approach, whether the builder is a brokerage, an advisory firm or an investment manager. If the market were served by a string of firm-specific assistants, each one would only have a slice of the industry's data and workflows: the mandates, geographies, asset types and processes of the firm that controls it. That is not a claim about JLL's models, the quality of JLL GPT, or the scale of the data JLL describes behind Falcon. It is our argument about what any single-firm build can see.

The risk with this approach, in our view, is that gaps and biases creep into the model when it mostly learns from one organisation's history. Not seeing the wider market, we think, leaves in-house "agent GPTs" capped out as OK, but not transformative. Again, that is our opinion of the in-house pattern, not a measured statement about JLL's performance.

There is also a trust question we think every investor, landowner or competing firm should ask. Can you rely on investment decisions and insights created by a model controlled by another market player, potentially with competing interests? We are not saying JLL would misuse a model, and we are not attributing a conflict to JLL's tools. We are arguing that control and incentives matter when an assistant sits inside one market participant, and that many firms will want their AI to run on their own data, under their own governance, with every source open to review.

Why we think specialist platforms beat in-house builds and generic assistants

Firms that cannot use JLL GPT, and do not want to spend years building an equivalent in-house stack, usually look at two other routes: a generic enterprise assistant such as ChatGPT Enterprise, Claude for Work or Microsoft Copilot, or a specialist commercial real estate platform. We think the specialist route is the one that holds up under investment committee scrutiny.

Generic assistants are easy to start, and the leading ones already work across your own documents: Microsoft Copilot, for example, is grounded in your organisation's Microsoft 365 data (see our Copilot comparison). In our view, what they are not built for is the real estate workflow itself. Reconciling an offering memorandum with the rent roll, the T-12 and the lease amendments, and turning that into a cited IC memo your committee can audit, is still work your analysts have to set up, prompt and check themselves.

An in-house build can be trained on a firm's own language and history, which is why JLL's direction is interesting. We still think most investment and asset management teams should not try to copy it. A foundation platform, data pipelines, security reviews, model operations and a library of internal tools are a multi-year programme. Even then, the assistant remains tied to one firm's slice of the market, and third parties are right to ask who controls it.

Fifth Dimension is the option we built for that gap. It is a decision intelligence platform for commercial real estate that connects the documents, data, systems and expertise your firm already relies on, from deal screening and underwriting to lease abstraction and portfolio monitoring. Your team stays in control of the judgement. Fifth Dimension takes on the work between receiving the information and making the decision.

That is a different design from both a generic chat seat and a single-firm in-house assistant:

  • Your data, your environment. Customer data remains private and is never used to train the models that power Fifth Dimension. Each customer operates in a separate product environment, with no data cross-pollination between clients. Fifth Dimension is SOC 2 Type II compliant, ISO 27001 certified and GDPR compliant, with US data kept in the US and EU data kept in the EU. On-premise or air-gapped deployment is available, alongside SAML single sign-on, role-based access, IP allow-listing and data lifecycle management. See Fifth Dimension security.
  • Connected to the systems you already run. Fifth Dimension connects Excel, SharePoint, Argus, Dealpath and more, bringing them into one intelligence layer across your stack, and integrates with Yardi. Workspaces give teams a shared place for an investment, asset or portfolio decision, synced with your underlying data rooms. See Fifth Dimension integrations.
  • An assistant that works the way your firm works. Fifth Dimension's AI agent understands your firm's terminology, processes and priorities, and applies them consistently, so screening notes, lease abstracts and IC memos follow the same institutional standard.
  • Evidence a committee can review. Evidence & Controls keeps sources, assumptions, review steps and changes visible, so your team can understand how an output was produced. See Fifth Dimension for investment managers.
  • Lease abstraction is part of the same platform. Lease review, amendments and side letters sit alongside underwriting and portfolio monitoring, not in a separate product line.
  • Clear commercial terms. Base starts from $139 per user per month, billed annually, with a $20,000 annual minimum. Enterprise starts from $50,000 a year. See Fifth Dimension pricing.

In short, we like that JLL invested in JLL GPT and JLL Falcon. We still like our approach more. We think a specialist platform that runs on your data, cites its sources, and is not controlled by another market player is the more durable answer for CRE deal and asset teams.

See it on your own workflows. Book a Fifth Dimension demo.

FREQUENTLY ASKED QUESTIONS.

What is JLL GPT?
JLL GPT is JLL's in-house generative AI assistant for commercial real estate, announced in August 2023 as a large language model built by and for CRE. It is also searched as JLLGPT, JLL AI and JLL ChatGPT.
Is JLL GPT the same as ChatGPT?
No. JLL GPT is JLL's own CRE assistant, not OpenAI's consumer ChatGPT product. JLL has said businesses were hesitant to use tools such as ChatGPT due to security concerns and inaccurate outputs. JLL does not disclose which underlying models it uses.
Is JLL GPT available to the public?
According to JLL's public information, JLL GPT was described as a tool for in-house use by JLL's global workforce. We have not found an official JLL page offering JLL GPT to the public or to clients as a standalone product. Clients may see AI through JLL's services and Falcon-powered tools.
What is JLL Falcon?
JLL Falcon is JLL's AI platform, launched on 29 October 2024. JLL says it powers the next iteration of JLL GPT and a wider set of JLL applications, and describes it as built on nearly two petabytes of JLL's own transaction and operations data.
When was JLL GPT launched?
JLL announced JLL GPT on 1 August 2023. Fifth Dimension first published this article on 31 July 2023 and has updated it as JLL has released later public information, including JLL Falcon in October 2024.
How many JLL employees use JLL GPT?
JLL said more than 47,000 professionals had used JLL GPT as of October 2024. In June 2026, JLL's CTO said more than 50,000 employees actively use it. Those figures come from JLL, not from our own measurement.
If we cannot use JLL GPT, what should we use instead?
We think the better option is a specialist CRE platform that connects to your own data rooms and systems, keeps every source visible, and is not controlled by another market player. Fifth Dimension is built for that job.

Tools & libraries.