When was the last time you asked your CRM vendor what happens to your data if you decide to leave?
Most agents have not. Not because they are not curious, but because the answer has always been understood at some level — and the understanding is uncomfortable. The lead histories, the landlord contact trails, the campaign performance records, the years of prospecting activity logged inside the platform — are difficult to extract, often incomplete in export, and in some cases contractually contested. Leaving means starting over. That is not a coincidence.
The CRM models that dominate estate agency technology were not designed around the agent's interests. They were designed around retention. Every architectural decision — proprietary data formats, limited export functionality, AI tools walled inside the platform, integrations that talk only to the vendor's own products — was optimised to raise the cost of switching. You are not a customer. You are a tenant. And the lease is written entirely in the landlord's favour.
How did CRM lock-in get worse when AI was added?
For years, the mechanism was straightforward: your data lives in our format, in our system. Extracting it costs time, money, and the loss of everything the platform built around it. The switching cost is high enough that most agents renew on terms they would reject if they were signing for the first time.
Then came AI. And the lock-in deepened considerably.
It is not just contact records that now live inside vendor ecosystems. It is the AI's learned understanding of your local market — the predictive models trained on your usage patterns, the outputs you acted on, the intelligence that accumulated silently as you used the tool. When an agent leaves a CRM with AI-powered features, they do not just lose their contacts. They leave behind the machine's accumulated knowledge of their patch. They start from zero again. That is not a side effect of AI adoption inside closed platforms. It was the design intent.
"When your intelligence lives inside a vendor's walls, in their format, you don't have data. You have access to data. There is an important difference."
The forward-thinking agents and developers who have recognised this pattern are asking a different question. Not which CRM has the best AI, but how to connect property data to ChatGPT and the other AI tools they already own — so the intelligence layer lives with them, not with the vendor. The answer is open data standards and API-first architecture. Specifically, a standard called the Model Context Protocol.
What does an open property intelligence platform look like?
Open is not a marketing claim. It is a specific set of technical and commercial commitments. Your data is available in standard formats without proprietary software. You can connect it to any tool — internal, external, AI or otherwise — without vendor permission. The AI outputs you generate are yours permanently. You can leave, completely and cleanly, and take everything with you.
In property technology, this is what a property MCP server architecture delivers. MCP — Model Context Protocol — is an open standard that allows AI models like ChatGPT or Claude to connect directly to live external data sources. When a property data provider operates an MCP-compatible server, an agent or developer can connect their AI tools to daily-refreshed sales data, landlord records, EPC information, and environmental risk data — without copying anything into spreadsheets, without waiting for the CRM to build an integration, without any vendor intermediating the connection.
The UK property data ChatGPT integration that would have required bespoke engineering a year ago now has a clean, documented architectural answer. And the agent owns every output it produces.
What questions should you ask before any PropTech contract?
Before the next CRM renewal arrives, before the next data platform demo, there are four questions worth asking — and worth insisting on clear, unqualified answers to.
- Can I export everything — every contact, every landlord record, every campaign result — in a usable format, right now, without paying for the privilege?
- Can I connect this data to tools outside your ecosystem without your permission or your involvement?
- What happens to the AI outputs I generate using your platform — do they remain accessible if I leave?
- Can I switch to a competing provider with two months' notice and take my data cleanly with me?
Vendors built on genuine openness answer all four immediately, because transparency about data ownership is their commercial advantage, not their vulnerability. Vendors who hedge, redirect, or bury the answers in terms and conditions are telling you exactly what you need to know.
The PropTech vendors who will matter in five years are the ones who understood this before their customers forced the conversation. The ones building closed ecosystems are building walls that market forces and regulators are already working to dismantle. The timeline is the only uncertainty.
📌 KEY TAKEAWAY
Your CRM is engineered around lock-in. Ask the four questions above before signing anything. propalt.ai is MCP-compatible, open-format, and built on the principle that your intelligence belongs to you — connect it to ChatGPT, Claude, or any tool you choose.
Sources
—Data (Use and Access) Act 2025 — legislation.gov.uk
