Green Street adds news and workflows to its real estate MCP server

Green Street, the commercial real estate research firm based in Newport Beach, announced on October 13 two expansions to its MCP server: the integration of Green Street News and a first batch of packaged workflows. It comes barely two months after the company declared the server generally available in mid August, with connections to corporate instances of Claude, ChatGPT and Gemini.

The figure behind the move is in its own documentation: the server provides access to more than 20,000 research reports, more than 150,000 news articles and more than 40 years of proprietary sector data. Until now, the news component was listed as "coming soon". With the announcement released today, available in full on PR Newswire, that is no longer the case.

What the expansion includes

The news tool lets users ask natural language questions about Green Street News coverage, which spans the United Kingdom, the European Union, Australia, Canada and the United States. Answers arrive synthesized and with a link to the full article on the Green Street platform. According to the company, these queries can be combined in a single conversation with the server's other datasets: Research, Company Data, Market Data, Forecasts and Sales Comps.

The second addition is workflows. Green Street presents them as structured views that go beyond single question retrieval and return the same format every time they run. The first two are called Ask Green Street Concierge and Get Green Street Digest, and the company promises to expand the library regularly. The release does not detail exactly what each one does or how they are implemented.

Richard Neale, the company's Chief Product Officer, sums it up: "Data is most powerful when paired with context. With News now part of the MCP Server, every prompt can draw on the stories behind the numbers."

Why it matters

What matters here is less the sector than the pattern. A paid data provider, with a product built over decades, decides that the channel to reach the analyst is no longer just its website or its API, but the assistant the client already has open. On the product page the catalogue is broad: automated asset valuation (AVM), transaction comparables, forecasts for cap rates, CPPI and occupancy by sector and market, and company profiles with NAV premium and target prices.

The workflows target a problem we often see in MCP projects: exposing tools does not guarantee that users know how to combine them. An analyst asking about the London office market can get very different answers depending on how the request is phrased. Packaging a fixed sequence of queries with a stable output format is how ten people on the same team end up with comparable reports. The protocol itself covers this idea with the prompts primitive, templates exposed by the server; Green Street does not say whether its workflows rely on it.

Some figures deserve caution. The company claims that beta participants completed in 15 minutes tasks that used to take several weeks. It is a vendor claim with no published methodology. More useful is its explicit commitment that client data is not used to train or fine tune models, something any compliance department will ask about before connecting the server. Its CTO, Travis Valentine, defends the approach with a hard to dispute argument: if AI reasons over unverified data, the result is answers that sound very competent but are wrong.

Who it is for

The direct audience is Green Street's institutional clients: funds, REITs, real asset managers and research teams that already pay for the subscription. For them, the update removes steps, for example when cross referencing a breaking story with forecasts for the affected market without leaving the conversation with Claude.

For technical teams in other sectors, the case works as a reference. It shows how a vertical provider organizes its server by dataset, adds a workflow layer to standardize usage and separates what is in production from what is still in beta, such as its research Q&A tool.

At ElephantPink we think packaged workflows will end up separating the data MCP servers that get used daily from the ones that only get installed. It remains to be seen how much flexibility they keep when an analyst needs to go off script.