Every major AI company in the world has agreed on a single way for AI agents to connect to external data, and Microsoft has already shipped it in Teams. Meanwhile, the UK property industry has been building its own open data standard, now live with Lloyds, Connells and LMS. Ed Molyneux explains why these two developments, arriving at the same time, will reshape the way conveyancing gets done over the next 12 to 18 months, and what it means for every knowledge-based profession after that.
You have probably read about AI agents over the past year. You may have used a chatbot in your own work. You almost certainly have not yet had to deal with an autonomous AI agent acting on behalf of your client’s buyer, or your lender, or the estate agent on the other side of the transaction.
But that is about to change, faster than most people in our industry expect.
A universal language for AI agents
In the last few months, something quiet but significant has happened. Anthropic, OpenAI and Google have all adopted the same open standard for how AI agents talk to external systems, called the Model Context Protocol, or MCP for short. Microsoft has already shipped MCP support inside Teams, the chat and collaboration tool sitting on millions of professional desktops.
Reliable reports now suggest Apple will be next, baking native MCP support directly into iOS and macOS and putting it on more than 2.5 billion devices, which means not just the iPhones and iPads people carry around, but also the Mac you might use at your desk and, in time, every modern operating system speaking the same language as the AI agents being built by every major AI company in the world.
This is the moment the AI agent era stops being a developer story and becomes a public infrastructure story.
You do not need to understand the technical details, only the shape of it. Until now, getting an AI to use an external system has been a hand-built, custom job, with every plug-in, connector and workflow wired up one at a time. MCP turns all of that into a single agreed standard.
An AI agent can ask a system, in a clean and predictable way, what data it holds, what actions it allows, and what the agent is permitted to see, and the system can answer. It is the difference between every appliance in your house needing a different plug, and every appliance using the same socket.
When that socket sits on billions of phones and computers, the implications for the way professional work gets done are very large indeed.
Property is unusually ready for this
If MCP is the universal socket, the natural next question is what it plugs into in our world, and that is where property is in an unusually strong position.
Property data is famously fragmented, with title information at HMLR, lender requirements in the Lenders’ Handbook, searches at local authorities, lease terms in PDFs and seller disclosures in TA forms, and even when an AI can read all of it, it has no easy way to know which bits are reliable, which are recent, and which were vouched by a qualified professional rather than typed in a hurry by the seller.
That is exactly the gap the Property Data Trust Framework, or PDTF, was built to close. The PDTF is the open standard the UK property industry has been quietly building together, and it is the same framework that underpins the digital home buying journey announced recently by Lloyds, Connells and LMS.
It gives every fact about a property a defined place, a defined type, and, crucially, a defined source, with every claim carrying provenance: who said it, when, with what evidence, and with what confidence.
On its own, the PDTF is powerful. With MCP layered on top of it, the two standards together become the unlock.
What changes in the next 18 months
An AI agent that knows the title was extracted from HMLR on a specific date, that the lease length was vouched by the conveyancer, and that the flood risk was self-reported by the seller, is no longer a chatbot.
It is a different kind of system, one that can act, defer, escalate and explain, because it genuinely understands the trustworthiness of its inputs.
That is how a buyer’s AI agent will be able to grapple with the whole picture of a property and tell them honestly whether it meets their needs, and how a lender’s AI agent will be able to assess risk against banking-grade data in real time before issuing an offer.
It is how the seventy per cent of conveyancing work that is intelligence, chasing and cross-referencing begins to happen automatically, leaving the thirty per cent that is genuine professional judgement firmly with the conveyancer where it belongs.
Why this is bigger than property
This is not only a property story, of course. The same shift is coming for every knowledge-based profession.
Accountants whose clients ask a bookkeeping AI agent to reconcile suppliers automatically. Insurance brokers whose customers’ AI agents pre-fill quotes from verified data. Doctors whose patients arrive with structured pre-consultation summaries assembled by an AI working on their behalf.
The pattern is the same in each case: a universal way for AI agents to connect, sitting on top of trustworthy, properly provenanced data.
The professions that get there first, with their own data infrastructure in good order, will spend the next decade automating the routine work and selling more of the hard, judgemental work. The professions that wait will spend it watching their best clients get used to a faster, better service somewhere else.
What it means for you
The good news for working conveyancers is that none of this requires you to install anything new, learn MCP, or become a developer.
What it does mean is worth thinking about over the year ahead: whether the data you produce on every transaction lives in formats an AI agent can actually read and trust, whether your firm’s own tooling can plug into the open standards now spreading across the rest of the industry, and whether your reports, your enquiries and your written work could be machine-readable as well as human-readable without losing any of their professional value.
The AI agent era in property does not arrive through a single product. It arrives through open standards that finally let trustworthy data and intelligent AI agents meet in the middle, and that meeting is now.
About the author
Ed Molyneux is co-founder and CTO of Moverly, the property intelligence platform working with LMS, Connells Group, and Lloyds Banking Group to bring structured, verified data to property transactions. He is the original author of the Property Data Trust Framework (PDTF), the open standard for machine-readable property data now adopted across the industry.
