Building “Joy”: A Custom AI Voice Agent

When Renjoy named their new AI voice agent “Joy,” it wasn’t just for show. It was about naming a feeling—what it means when your systems work together, when the answers are already there, and when the technology feels like part of the team. There’s a kind of delight in watching AI perform not as a novelty, but as something reliable, responsive, and fully integrated into the day-to-day.

That kind of AI wasn’t possible without the groundwork. Renjoy had already built the infrastructure—where operational knowledge was accessible, editable, and portable. So when capabilities like Conduit's voice agent came into the picture, they weren’t starting from scratch. They were ready to give their systems a voice. And they called it Joy.

The Challenge

Property managers often spend years enriching tools with custom data—then lose it all when switching vendors. Renjoy had been through this cycle before, jumping between cleaning platforms like Turno, EasyCare, Aperto, and Breezeway. Each time, valuable context got lost.

“We always wanted the option to jump platforms,” said Michael Vialpando, Co-Founder & CTO at Renjoy. “Too many managers enrich their data in a tool and then get stuck. We wanted to avoid that entirely.”

To break the cycle, they invested in their own Airtable infrastructure, using it as a neutral source of truth. This gave them flexibility when adopting new tools, including Conduit. It also made them faster. Implementation was smoother. New features could pull from a consistent dataset. Their systems stopped competing and started cooperating.

The Solution

Renjoy left Akia after realizing the product wasn’t keeping up with where the industry was going. They didn’t want tools that simply bolted on AI—they wanted platforms built around it. “The big thing was just not hearing anything,” said Vialpando. “We were using AI everywhere in our business. Our vendors weren’t.”

Conduit gave them that, and within weeks, they reached over 40% automation on inbound guest messaging. "You guys just move fast and build well," said Vialpando. "It’s what we were looking for in a vendor—someone putting AI at the forefront, not treating it like a feature buried in the backlog.”

The next step? Joy—their AI voice agent. Powered by their structured property data, Joy is designed to handle both inbound calls with guests and sales demos with new owners. It doesn’t just sound smart—it is. When a guest calls and hears parking instructions that reference landmarks and access codes, the experience feels human, not robotic.

“We want Joy to prove herself immediately. Tell a guest to look for the blue house and enter through the gate—that’s how you show it’s not a gimmick,” he said.

Key results from using Conduit:

• Achieved 40%+ message automation via autopilot in the first week
• Launched "Joy," a voice agent powered by their property data
• Reduced team workload through AI-suggested replies and message handling
• Accelerated onboarding and implementation thanks to structured internal data
• Built team confidence by integrating knowledge directly into messaging workflows

A Smarter Setup

Renjoy’s internal Airtable setup allowed them to:

• Onboard to Conduit without rebuilding information.
• Move between platforms without losing operational knowledge.
• Give their team confidence that their input won’t disappear with the next switch.

“You stop being afraid to teach the system when you know the data’s yours,” said Vialpando.

The Impact: Fewer Agents, Higher Leverage

With Conduit, Renjoy has fewer communication staff—but they’re working at a higher level. AI takes care of the repetitive work. Suggested replies speed things up. Teaching the AI happens naturally during support workflows, not in side spreadsheets.

“We tell our team: you’re not here to type the same answers over and over. You’re here to solve higher-level issues,” Vialpando explained. “That means we can grow without hiring dozens of people just to manage conversations.”

The outcome isn’t about reducing headcount. It’s about creating room for higher-leverage work. The people who stay aren’t stuck answering the same five questions—they’re solving the edge cases that matter.

Looking Ahead

Renjoy plans to entere new regions using the same playbook: tight backend systems, flexible data structures, and AI trained on the right information. Once those pieces are in place, localization becomes the only real challenge.

“We don’t want one whale that owns everything,” said Vialpando. “We want to help thousands of people reach financial freedom. That means we have to get the backend right—then replicate it.”

They’re not trying to build the flashiest front end. They’re building a company that can scale without falling apart. And Joy—along with the infrastructure behind it—is a big part of how they plan to do it.