Quick summary:
- West Shore Home does 55,000 projects a year across 43 locations and uses machine learning to match sales reps to appointments the way baseball teams match hitters to pitching situations
- The company’s obsessive data centralization and W-2 employee model gave it a structural advantage that made AI adoption feel like a natural next step
- While 95% of AI pilots fail to reach production, West Shore Home sidesteps the failure pattern by framing every initiative around employee quality of life and starting small enough to earn organizational trust
West Shore Home’s sales managers used to set the daily board by hand. They’d look at a list of appointments, eyeball which rep should run which job, and try to minimize windshield time while matching skills to projects. It worked well enough. Then Eppie Vojt showed them what “well enough” was costing them.
Vojt, West Shore Home’s Chief AI and Digital Officer, borrowed a concept from baseball — wins above replacement (WAR) — and built a machine learning model that forecasts the expected value of every appointment based on dozens of variables no human manager could hold in their head at once.
Sales reps in pilot markets now experience 10% to 20% less drive time, better yield per appointment, and a scheduling engine that works at a scale no whiteboard ever could.
Speaking at HumanX in San Francisco on this episode of The Disruption Is Now, Vojt tells host Greg Matusky how a “lunchbox business” became one of the most compelling AI case studies in the country, doing $1 billion in home renovations with 43 locations in 24 states, and what other organizations keep getting wrong.
Watch now:
Key takeaways
Moneyball for home renovation matches reps to jobs nobody would have assigned them
Baseball organizations often evaluate players by how many wins they produce above a generic replacement. Vojt adapted the same logic for West Shore Home’s sales force.
Instead of measuring reps on raw conversion rates or total net sales — the industry standard — his team built a machine learning algorithm that forecasts the expected net sales value of every appointment based on attributes a manager can’t easily weigh. Homeowner demographics, product category, day of week, time of day, and the lag between scheduling and the visit all factor in.
The model reveals granular strengths that intuition misses. One rep might be a killer on evening shower appointments but struggle with morning door calls. Another might convert at a high rate in specific geographies. The system matches these patterns against incoming appointments and routes accordingly.
The biggest challenge was the sales managers. Setting the board had been their responsibility, and they’d done it well enough with the information available to them. Vojt’s team won buy-in by leading with drive-time reduction, something managers could feel immediately, rather than yield optimization, which takes longer to prove out. The company is running a randomized control study across branches before a full rollout.
Years of boring data work gave West Shore Home a head start most companies don’t have
CEO B.J. Werzyn built West Shore Home with obsessive data centralization in mind.
Every acquisition, every new branch, every operational change consolidated into a single source of record. When the AI moment arrived, the company had the foundation that made everything else possible.
When a sales consultant closes a deal in someone’s living room, the customer can schedule the installation date on the spot. That’s possible because the system pulls permitting information, inventory data, and installer availability in real time. Vojt described the overlap of those three datasets as a Venn diagram that enables instant scheduling. A subcontractor model, where availability data lives outside the company’s systems, could never do that.
The same data infrastructure now feeds the Moneyball algorithm and the company’s performance marketing. Vojt’s team sends differentiated lead-value signals back to platforms like Meta and Google, telling them which types of leads produce the highest expected net sales. The machine learning model that scores appointments also scores leads, creating a feedback loop that makes ad spend more efficient without requiring anyone to wait for a completed sale before measuring results.
95% of AI pilots fail because nobody earns trust before scaling
At the HumanX keynote, Vojt heard that 95% of AI pilots fail to reach production. A separate finding showed 56% of CEOs surveyed said they were getting nothing from their AI investments. He sees the root cause as a broken chain of command. Boards demand an AI strategy from C-suites. C-suites demand one from line managers. What comes back is box-checking.
West Shore Home avoids this trap by constraining what Vojt calls the “blast radius.” The point-of-sale scheduling AI launched with one brand and one product line. If something broke, the damage was contained. When it worked, the organization gained confidence.
Vojt framed it as a trust-bank problem.
Early wins make deposits. Early failures make withdrawals. Organizations that try to go big on a first AI initiative and stumble find it nearly impossible to get a yes on the next one. He described the inverse approach, where leaders rush an implementation or assume anything less than a full-percentage implementation will be perfect, as a guaranteed way to destroy trust rather than accelerate it.
Framing AI as a cost-cutting tool guarantees resistance
West Shore Home has never approached AI through a headcount savings lens. Every initiative maps to the company’s mission of bringing happiness to every home. That might sound like marketing, but it changes how the entire organization receives new tools. For example, when marketing gets more efficient, the company doesn’t cut the team. It generates more leads, which means it needs to hire more design consultants and more installers to fill the pipeline.
There’s also a subtler problem. Many organizations are so competitive internally that employees sabotage AI adoption. They find every flaw, seize on hallucinations, and use imperfections to discredit the effort, applying a standard of perfection to AI that they'd never apply to a Google search they've been verifying for years.
Vojt’s team sidesteps this by embedding AI so deeply that employees experience the benefit before they have a chance to build resistance. The salespeople whose schedules are now optimized by machine learning don’t even know AI is making the assignments. They just know the schedule works better than it did last month.
The challenger brand mindset might be the real secret to AI adoption
Matusky asked Vojt whether West Shore Home sees itself as a dominant brand or a challenger, given that big names like Jacuzzi now compete in the same one-day-install space. Vojt said the company will always maintain an underdog mindset. In fact, he sees complacency as the bigger risk than competition.
That matters for AI adoption because large, established companies tend to struggle more than nimble ones. They carry sunk costs in legacy systems and institutional inertia that makes change painful. Meanwhile, West Shore Home started from a single retail storefront in a small Pennsylvania town. It didn’t inherit legacy infrastructure. And its leadership has made “get better every day” part of the hiring criteria, selecting people who match the challenger mentality.
Matusky drew a connection to what AI leaders often say about the future belonging to companies that stay hungry. West Shore Home has already crossed the billion-dollar revenue mark with a lean AI team of eight. It keeps growing because it treats AI as a growth engine. This has helped the company hit record performance across most metrics this year, even while competitors in the same industry blamed the market for declining results.
Key moments
- What a home renovation business is doing at an AI conference (0:27)
- How point-of-sale scheduling uses the Venn diagram of permits, inventory, and installer availability (4:39)
- The Moneyball concept adapted for sales rep optimization (6:33)
- Selling AI to skeptical sales managers (9:17)
- How AI infrastructure makes acquisition targets more compelling (11:36)
- The growth playbook for greenfields, acquisitions, and new service lines (13:55)
- How machine learning signals feed back into Meta and Google ad platforms (16:35)
- 95% of AI pilots fail to reach production, and why (21:27)
- The internal sabotage problem and the Google double standard (24:10)
- Why “learn how to learn” is Vojt’s advice to his kids (33:43)
- Which legal jobs AI can already do better than humans (39:52)
Q&A with Eppie Vojt, Chief AI and Digital Officer at West Shore Home
Q: West Shore Home is over $1 billion in revenue. Where does growth come from next?
A: We have a lot of optionality for how the business grows. We could grow organically through improved performance, and we’re pulling those levers right now. AI implementation on our marketing funnel means that we can grow and seek more appointments with the same spend. We can also introduce new service lines.
There are plenty of things that fit the model of something we could be in and out of in one to three days. We’ve done that already by expanding into full bathroom remodeling and full home flooring. And we can grow into new territories. We just launched greenfields in Boston, Nashville, and Huntsville, Alabama.
Q: How does your AI platform change the conversation with acquisition targets?
A: I think there’s a compelling story that we can tell where AI enables a trajectory for the business going forward that didn’t exist otherwise.
For owners of those businesses, they still care about their teams. They want to know that people are going to have a good experience transitioning into West Shore Home — less confusion, less chaos. On the AI side, some sellers might have an interest in rolling equity into the business rather than exclusively taking cash, because rolling some capital into a company with this kind of trajectory could be massively valuable.
Q: What AI trend do you think is most overhyped right now?
A: AI voice is one that’s got a lot of hype that I don’t entirely believe in yet. There’s still a lingering perception from any time someone has a non-human interaction on the phone; people are scarred from years of terrible IVR experiences. They have this innate reaction to it that it’s bad, even if it’s really good now.
Q: What’s the first thing you’d automate in any business?
A: Anything that has audits in the process. If you’re running manual audits and compliance checks, those are the first things to go.
Q: What job disappears in five years?
A: Tier-one voice work. I think customer service phone interactions will go entirely to AI, and it’ll be exclusively escalations beyond what an AI agent can handle. My oldest son is an electrician, and he decided to go down that path. I think that becomes increasingly valuable. We’re going to need a lot of electricians for these data centers.
Q: What if AI development stopped tomorrow?
A: Our business still has so much potential to implement this capability. If this were the best AI would ever get, I think we’d have ten years of optimization just in getting the algorithms to be more elegant and more efficient.
