Every call went through one front desk. Now they are building an AI receptionist.
Cortes Management Group is a family-operated real estate agency in Cockburn Central. It looks after residential, commercial and strata property and sells property in Perth. Its Managing Director, Cristian, came to our one-day AI workshop with a plan already mapped out: an AI receptionist that answers tenant and owner calls. The build is in progress, one tested step at a time, with go-live targeted for early 2027.
What is Cortes Management Group building?
An AI receptionist. A voice agent will answer the office phone line, work out what the caller needs and put them through to the right team. It is being built now and is not live yet.
Cristian described the finished version on our first call. An owner rings to ask whether their tenant has paid the rent. The receptionist asks for the property address, checks the property management system and answers. If the question needs a person, it puts the caller through to the property manager, and a short email summary of the call is already waiting for them when they pick up.
That is the end goal. Getting there means answering the phone well first, then connecting the data behind it.
Why did they want an AI receptionist?
Because every call landed on one desk, and the answers callers need sit in separate systems. Cristian wanted fewer calls coming in and out of the office, and hours handed back to his team.
The agency was not new to AI. The whole team already used Gemini across Google Workspace, with custom assistants set up for marketing, sales, strata, commercial and residential work. What was missing was a link to the systems the work actually lives in: PropertyMe for residential, with separate software for strata and commercial. Cristian wanted AI to reach that data directly, "without the staff member having to log into the software, download the reports, then upload it to Gemini and work from there."
The goal that I've set for myself is to be able to get this implemented by January, February next year. I want this all up and running.
Where did our workshop fit in?
It was the starting point, not the whole build. On our first call we told Cristian straight that a voice agent was more than a one-day job. But Cristian had already mapped out what he wanted, so the plan became: learn Claude at the workshop, get the foundations in, then build the receptionist with our support.
Before the day, we ran a setup call to get Claude installed and his email connected. On the day, Cristian made a start on the voice agent while everyone in the room worked on their own business. Since then he has used our follow-up support to work through the phone setup and the voice agent with us. The team has also started connecting Claude to their accounting software.
How is the receptionist being built?
In stages, with each one tested before the next. The first stage is the phone line itself: a new number answers first, the voice agent takes the call, and anything that needs a person goes through to the team's existing phones.
The first design had to change. The office phone platform could not hand the agency's own greeting and hold music to an outside service, so the new number became the front door instead: it plays the greeting, hands the caller to the voice agent, and passes anyone who needs a person back to the existing phones, where a property manager picks up as normal.
The switch-over will be tested before it is made permanent, so callers never lose the agency's greeting along the way.
Where it stands in mid-September 2026:
- The call flow is mapped out: the new number answers, the voice agent handles the call, and a person takes over when needed
- The new phone number has been approved
- A first demo of the voice agent has been built and shared for testing
- Next: deciding where the new number shows for customers, then setting it up together
What will it do once it is live?
Answer and sort calls, look up live details while the caller is on the line, and hand over to a person with a summary. The plan also includes calls out to tenants: reminders about unpaid rent, upcoming inspections, maintenance updates and leases coming to an end.
Cristian wants those calls to run seven days a week. If a tenant says on a weekend that they are in financial difficulty, the agent is meant to capture what is going on and email the property manager, so it is waiting for them when they walk in on Monday. During the week, those calls go straight through to a property manager.
The live data is the hard part, and it is the second stage. We have mapped out the options for getting the property systems' information in front of the receptionist, and the choice comes once the phone side is proven.
Why build it slowly?
Because a voice agent that gets it wrong is speaking for your brand. Cristian was clear from the start that he wanted it tested properly before it goes live. In his words: "I am very committed to get it done. I'm not in a rush to implement it."
He has received hundreds of AI calls himself, and knows how quickly a bad one frustrates people. As he put it, "I would hate any of my clients or prospective clients to go through the same thing." Building in stages means each part is proven before the next one depends on it.
Cristian walked in with the whole thing mapped out. A voice agent is quick to build and slow to get right, because the testing is the real work. So we are going in order: the phone line first, then the live data behind it.
| Step | Before | Now |
|---|---|---|
| Answering calls | Every call goes to the receptionist | Same today, with a voice agent being built to answer first |
| Phone line | One phone platform takes every call | A new number approved to answer first and pass calls through |
| Report data | Downloaded and uploaded to AI by hand | Still by hand, with live lookups planned for the receptionist |
| AI tools | Gemini assistants for each team | Claude set up alongside them, with email connected |
| Tenant reminder calls | Made by property managers | Mapped out for the voice agent, not built yet |
Could this work for other property managers?
Yes. Plenty of agencies look the same: calls landing on one desk, and the answers sitting in property software that does not connect to anything else. The order Cristian chose works for any of them: learn the tools, answer the phone well, then connect the data.
We cover where AI fits an agency in AI for real estate, and voice agents in chat and voice. If you would rather build it yourself, like Cristian, start with our AI training.
Common questions
Is the AI receptionist live yet?
Not yet. It is being built in stages. The new phone number has been approved and a first demo of the voice agent has been shared for testing. Cortes Management Group is aiming to go live in January or February 2027.
Will callers still be able to reach a person?
Yes. The plan is for the voice agent to answer and sort calls, then put anything that needs a person through to the right team. The call reaches the office phones like any other call, and the aim is for the property manager to have a written summary before they pick up.
Who is building it?
Cristian, the agency's Managing Director, with our support. He started at our one-day AI workshop and has used our follow-up support to work through the phone setup and the voice agent with us.
Will the AI call tenants too?
That is part of the plan, but it is not built yet. The idea is reminder calls about unpaid rent, upcoming inspections, maintenance updates and leases coming to an end. Anything sensitive, like a tenant in financial difficulty, goes to a property manager.
Updated 17 September 2026
More of our work
Does every call land on one desk?
If your team loses hours to phone calls and to copying reports between systems, we can help you plan the fix and build it in the right order. Tell us about it, in plain language, with no pitch.