How Haven Vacation Rentals Hit 90 Percent Automation and Got Airbnb Knocking on Their Door (2026 Update)

TL;DR

• Reached 90 percent automation across 270 properties, up from 40 percent in a few months
• 80 percent reduction in support staffing needs—from 25 people to two per shift
• 3,700 Google reviews at 4.9 stars, with 55 percent of listings earning Airbnb Guest Favorite status
• Quality scores held or improved throughout the automation push
• Team members evolved into conversational engineers, building workflows and training the AI

The Challenge

Haven Vacation Rentals manages roughly 270 short-term rental properties across Tennessee's Smoky Mountains. At its peak, the company had around 25 people in guest support, split across morning, evening, and weekend shifts. Five or six per shift, fielding every check-in question, lockout, cleaning issue, and complaint that came through.

Jack, Haven's CEO, watched the volume climb with every property added. Growth meant more messages, more calls, and more staff. The economics worked, but the model didn't scale without proportional headcount.

Haven had Conduit installed but was barely using it. Listings were uploaded, general information was filled in, and the system ran at about 40 percent automation with no workflows and minimal knowledge base work. Enough to take the edge off, not enough to change how the team operated.

Jack and his partner looked at the year ahead and realized they couldn't keep up at the current pace.

"We had a really big year in 2025, so I was like, I can't do this," Jack said. "And my partner was like, we're not gonna be able to do this ourselves."

The Solution

Haven's approach started with the team, not the technology.

Jack gathered the support staff for a meeting. He drew a hard line: their jobs were safe, and the goal was to make the AI as good as possible.

"I made a very hard line in the sand," Jack said. "Your new job is gonna get as good as this as possible. We're not gonna lay anyone off. That's not happening."

The first meeting didn't land. The team didn't believe him. He held two more before people started to buy in.

Once they had trust, Jack tied a incentives to the automation rate. Jonathan, the guest services manager, became the central person teaching the AI. Having one source of truth mattered because different team members handled the same situations in different ways. A door lock issue at 8 AM got a different response than one at 4 PM. Jonathan standardized those answers.

The team created a Slack channel called AI Avengers. Anytime the AI handled something poorly, people posted screenshots. Jonathan and Jack would respond with guidance on whether to teach the AI or adjust a workflow.

The push from 40 to 80 percent automation took a couple of months. Getting from 80 to 90 required more creative problem-solving, but Haven held quality scores at 4.8 or 4.9 the entire time. The incentive was tied to both metrics: automation rate had to climb, and scores couldn't drop.

The Transformation

At 90 percent automation, Haven's support workload inverted. The AI handled routine guest communication across 270 properties. The team focused on guests with real problems.

Jack saw it during a snowstorm. Ninety guests checked in the same day the Smokies got eight inches of snow. People were stuck. They couldn't get in or out. The automation rate held steady, and only three issues needed a human.

"This happened before," Jack said. "I remember looking at my phone, and I'm like, how in the world am I gonna look at all 100 of these problems? Then on Sunday, I looked, and there was three things. And I was like, oh, this is great. We can solve three."

The team that used to copy-paste door codes now builds workflows and trains the AI. Jack gives them problems to solve. One recent example: every time a guest brings a dog to a specific set of properties, the system needs to notify the cleaning company and push the information back. Jonathan built the workflow in about an hour, figuring it out through trial and error.

"That's how we've trained," Jack said. "Give them these problems. They just went and built the workflow and figured it out."

Haven also rethought how the AI communicates during gaps in its knowledge. Instead of stalling or guessing, the AI tells the guest it needs a moment to check with the team and will follow up within two or three minutes. Then it runs the workflow. Jack credits that small change with building guest trust, even when the AI doesn't have an immediate answer.

The staffing math changed permanently. Haven went from needing around 25 support staff to running with two people per shift. Nobody was laid off because of the AI. Some left for other reasons. Others moved into different roles. As more listings come on, Jack pulls people further into conversational engineering rather than hiring new support reps.

"We added 14 listings in January," Jack said. "We haven't had to add anyone."

Results

• 90 percent automation rate across 270 properties, up from 40 percent
• 80 percent reduction in support staffing needs—from 25 people to two per shift
• 59 percent of listings earned Airbnb Guest Favorite status
• 14 listings added in January without hiring additional support staff
• Maintained super host and premier host status throughout the automation push
• Team members transitioned from reactive support to workflow building and AI training

Looking Ahead

Haven is rebuilding its Slack triage workflows and expanding how the team uses Conduit's workflow engine. Jack is also exploring homeowner acquisition workflows with voice AI for inbound phone leads.

The knowledge base work continues. Jack's advice: don't try to finish it in two weeks.

"You can get automation quickly, but I think where people make a mistake is, the next two weeks, we're just grinding on this, and we're gonna be done," Jack said. "Do a two-week sprint, but if you really kinda just pick away at it for six months and just think, hey, we don't need to rush into anything, I think over time, it's just good to teach it and go on."

Haven Vacation Rentals operates roughly 270 short-term rental properties in Tennessee's Smoky Mountains. The company has been a Conduit customer for approximately two years. Jack is the CEO.

‍