labs@atheron:~/industries/real-estate/ai$ train agent --grounded

AI for buildings

A building assistant that answers residents.

Most resident questions have an answer in a document the manager already wrote. A frontier model can answer them from those documents and act on simple requests, which makes it the fastest and cheapest route here, provided resident data stays out of the prompt and anything contentious goes to a person.

Start a project See the example projects

// what matters here

What is different in this industry.

Answers from your building

The assistant answers only from each building's rules, notices and FAQs, and says when it does not know rather than guessing.

Does the simple things

It can open a repair request or book an amenity, using the same rules as the app, so a question becomes a ticket without a phone call.

People handle the hard things

Leases, disputes, payments and complaints go to the manager with the conversation attached. The assistant does not interpret a lease.

Resident data kept to a minimum

Names and unit details are left out of what the model sees unless a task needs them, and the provider's data terms are agreed before launch.

// example projects

Priced examples for this service.

Each one is hypothetical, labelled as such, and priced live from our rate card. Open any of them in the estimator and make it yours.

Example projectWeb appAIFrontier model

Building assistant for tenants

Hypothetical. Not a client, not a result.

// The problem

Front desks and managers answer the same questions all day about parking, garbage day, amenity rules and how to report a repair.

// What we would build

An assistant inside the resident app that answers from the building's own rules and notices and can open a repair request or book an amenity, on a frontier model because the source documents are not sensitive and it is the quickest route to launch.

// What is in it

Answers grounded in each building's rules, notices and FAQs
Tools to open a repair request or book an amenity from the conversation
Hand-off to the manager for anything about a lease, a dispute or a payment
Resident names and unit details kept out of the prompt where they are not needed
An evaluation set of real resident questions, run before every change

// Stack

Frontier model API
PostgreSQL with pgvector
Next.js
DigitalOcean

// estimate

Build: ≈ US$92,700 to US$142,000, delivered within 21 weeksCAD 132,000 to 201,800
Hosting: ≈ US$558 a monthCAD 795 a month
Support: ≈ US$1,000 a monthCAD 1,425 a month

Prices in your currency are estimates from today's Bank of Canada rate. All invoicing is in CAD or USD.

AI route: Frontier model
Model running cost: ≈ US$383 a monthCAD 545 a month

Timeline by milestone

Discovery: 2.9 to 3.4 weeks
Specification and evaluation plan: 0.4 weeks
Design approved: 2.4 to 3.4 weeks
Core features: 1.4 to 1.9 weeks
Full build: 0.9 to 1.4 weeks
Working pilot: 3.9 to 6.4 weeks
Testing and fixes: 0.9 to 1.9 weeks
Launch: 0.9 to 1.4 weeks
Production: 0.9 to 1.4 weeks

Outside our hands, and added to the calendar

API access from your property system vendor: 1 to 4 weeks

How it is paid

Deposit 20%: CAD 26,400 to 40,360
Discovery 2%: CAD 2,640 to 4,036
Specification and evaluation plan 10.2%: CAD 13,464.00 to 20,583.60
Design approved 2%: CAD 2,640 to 4,036
Core features 6.2%: CAD 8,184.00 to 12,511.60
Full build 4.1%: CAD 5,412.00 to 8,273.80
Working pilot 30.7%: CAD 40,524.00 to 61,952.60
Testing and fixes 2%: CAD 2,640 to 4,036
Launch 2%: CAD 2,640 to 4,036
Production 10.8%: CAD 14,256.00 to 21,794.40
Holdback, 30 days after launch (10%): CAD 13,200 to 20,180

Start from this example

// questions

Questions we get about this.

Why a frontier model rather than one we host?

The documents are not sensitive and the volume is modest, so a frontier model is the cheapest and quickest to run. If residents' personal data later needs to be in scope, an open-source model in your cloud is the next step.

Will it answer tenancy law questions?

No. It answers from your building documents and refers tenancy questions to the manager or the provincial tenancy office.

How is the building assistant kept up to date?

When a manager posts a notice or edits a rule in the portal, the assistant's source is updated with it. There is no separate content to maintain.

Back to the industry: Real estate and proptech About the service: AI agents and models

// next

Not quite your project?

Tell us what you have in mind. We will come back to you with a range and the questions that would narrow it. Or book a call and talk it through.

Start a project Book a call