At Salesforce’s Agentforce World Tour in Sydney last week, Geocon CIO Andy Magee delivered a big red warning: the success of agentic AI isn’t about rolling it out – it’s about testing it in the real world first.
He believes if you implement it blindly, you risk a combination of customer frustration and lost trust.
Magee, who is in charge of Geocon’s AI initiatives, told B&T that he wants every customer interaction to feel “unbelievably helpful,” due to the agent “always getting it right”.
The Australian developer is preparing to launch a new AI sales development agent in Q1, building on what Magee describes as the “massive impact” of its first production AI agent in customer service. Yet he remains acutely aware of the hurdles, from initial skepticism to outright resistance.
“We’re definitely not rushing this,” Magee told B&T. “You don’t really get to take it back once it’s out there. I’m confident we’ll do this right, and that’s why we’re taking the time—it’s critical when you’re taking this inquiry away from a human initially.”
No more bottlenecks
Geocon’s first AI agent was built to tackle “a long-standing operational blind spot”: defect management for tenants and property managers.
Around 60 per cent of Geocon’s apartment sales go to investors, meaning the end resident is often not the original buyer. When defects are lodged — anything from cabinetry issues to water leaks — the developer still holds responsibility during the defect liability period, but historically lacked direct visibility and structured engagement with tenants.
“The problem was scale”, Magee said.
“When a high-rise precinct settles, sometimes hundreds of apartments at once, defect requests surge overnight.”
According to Magee, human teams “struggled to keep up”.
“Tickets were often incomplete, missing photos or critical information, subcontractors sent jobs back and tenants waited weeks before anything progressed,” he said.
“We have some critical things that it looks at, for instance a leaky tap. Anything that indicates there could be leaking water, we have an internal requirement to make that an urgent case, because no one wants a leak in their apartment and also there’s potential further damage to neighbouring apartments.”
So we’ve got some context built into the agent, where it’s able to understand certain things and certain cues to ensure that we’re able to actually prioritise those kinds of problems.”
Magee said the idea of using an AI agent has proved successful.
Embedded on a public-facing page, the conversational bot builds a tenant profile from scratch – categorises the issue, ensures minimum data requirements are met and instantly routes the case to the correct subcontractor.
Urgent cues – such as those water leaks – trigger internal prioritisation rules. The system also checks whether the defect falls within the liability period, filtering out invalid claims and educating tenants in real time.
Magee said “what previously took weeks of back-and-forth now happens within a single session.”
The impact has been felt across all sides: tenants receive faster fixes, property managers can bulk lodge issues efficiently, and subcontractors receive complete, actionable tickets.
“It’s infinitely scalable,” Magee said. “When inquiry goes up, AI just handles it.”
Another agent on the way, this time in sales
Now Geocon is moving further up the funnel.
Its next AI agent, which is currently in advanced testing, will act as a sales development representative, intercepting property enquiries from websites, listing portals and social media before they reach external sales agencies.
“If someone enquires on a Sunday night, within five minutes they’ll hear from us,” Magee explained.
Rather than simply sending a brochure, the AI builds a contextual profile which understands budget, investor status and interests, while also answering lifestyle-based questions using curated local data on schools, cafés, restaurants and amenities near each development.
He said the goal is not to replace the human agents, but to just elevate them.
“No real estate agent wants to do data capture and grooming,” he said. “They want to close sales. If we can hand over a fully profiled, sales-ready prospect, that’s better for everyone.”
Crucially, the AI is designed with guardrails. It introduces itself transparently, identifies when a prospect wants to speak to a human, and automatically escalates if engagement stalls. Human oversight capabilities are also on the roadmap.
Why you can’t ignore the frustration
Magee is candid about a core tension: not every customer wants to interact with AI.
Older generations may resist. Some buyers will simply prefer a human voice. Others may assume AI adds friction rather than value.
His view is pragmatic.
“Everyone’s going to have to become more accepting of it,” he said, noting that AI is rapidly becoming the default customer service layer across industries.
But acceptance, he argues, “must be earned”. Poorly implemented AI lowers expectations, and well-architected AI changes minds.
The focus, therefore, is not on deploying technology for its own sake – something Magee believes many organisations get wrong – but on solving defined business problems first. Geocon spent years strengthening its CRM data governance before layering AI on top, ensuring the agents operate with meaningful context rather than guesswork.
“If you don’t have good data, it’s very hard to have AI with context,” he said.
For customers who disengage, the system quickly routes enquiries to humans. For those who embrace the interaction, the goal is a “wow factor” — an experience so helpful that it enhances trust rather than erodes it.
“I want people to say, ‘I dealt with Geocon’s AI and it was unbelievably helpful.’”
A high-stakes launch
The sales AI is expected to go live by the end of Q1, though Magee is deliberately resisting pressure to accelerate the timeline.
The stakes are high internally as well as externally. A misstep could undermine sales team confidence and damage brand perception. A successful launch, however, could redefine how property developers manage global agency networks and customer journeys at scale.
For Magee, the strategy is clear: control the narrative, nurture the customer and use AI to enhance — not gatekeep — human expertise.
In a sector often criticised for patchy service and inconsistent communication, Geocon is betting that intelligent automation, done properly, can become a competitive advantage.
Whether customers love it or hate it will soon become clear. But one thing is certain: AI agents in property are moving from pilot projects to the front line.
