Housing / FarmTable

Town of Vail Housing

Which restriction belongs to this property?

A property lookup brings deeds, restrictions and review history together. When a deed points to the wrong home, staff have a place to investigate.

The FarmTable AI Lab

Pull up a chair.

At FarmTable, we start with a question. What’s one part of your work that takes longer than it should?

Let’s take one property in Vail.

A housing specialist takes a call. Does this home have a deed restriction? In Vail, the deed might be in Laserfiche, while the housing team keeps its own property record. We started by connecting the places staff already look.

Which systems?

Laserfiche, OpenGov and Microsoft 365, including SharePoint and OneDrive. The housing CRM and GIS supply the property records we match against.

Read the fine print. Keep the page.

We pull deeds, covenants and housing guidelines securely from Laserfiche and turn scans into text. In this example, AI picks out a parcel number and the clause requiring the owner to live in the home. Each detail stays tied to the page it came from.

Same home. Different names on paper.

A deed can name an earlier owner. We match its clues against the parcel map, owner records and housing CRM. The parcel number ties this deed to the home even when the owner’s name no longer matches.

The slightly nerdy bit

A parcel, a person and a deed are different things. The ontology describes how they relate. An owner can change while the restriction still applies to the parcel.

Now ask about the home.

“Which restriction applies to this home?” In Groundwork, staff can ask in chat. The tools follow the parcel to its deed. Here, the answer is “owner occupancy required,” with page two ready to open. An annual review can use that same lookup.

Under the hood

MCP servers let the assistant call specific tools, like looking up a parcel or retrieving a document. Those tools give it actual housing records to work with.

The surprise was what didn’t line up.

Linking the records surfaced missing properties, deeds attached to the wrong home and errors in county records. For an audit or annual check, staff can spend their time on those exceptions.

How we test a revision

In our lab, we’re building a development swarm on FarmTable’s NVIDIA and HPE hardware. One agent can build a change while another tests an edge case. Our team reviews the results before bringing a revision back to the people who use it. Visit the AI Lab →

What would you put on the table?

That’s the Vail file. Open another folder to see what else our team brings to the table.

When the outcome is clear, we can quote a fixed bid. For ongoing work, we can join your team on an hourly basis.

Still figuring it out? Tell us where you get stuck.