When you're 2,000 customer service cases behind, the fastest fix isn't always the right one. That was the question facing Bryant Blair, CFO of Cambridge Air Solutions, when he weighed AI vs. human customer service for his team. AI could close tickets fast. But could it turn a frustrated caller into a raving fan?
We recently sat down with Bryant for a Hire Bloom AMA to hear how he answered that — and a lot of other questions every growing operator is sitting with right now. Here's what he told us.
Cambridge Air builds industrial heating, cooling, and ventilation systems. Bryant has spent over a decade in HVAC, and three years as CFO. The company's value proposition is simple: be "the easiest company to do business with in the space."
That promise got hard during busy season. "We've been inundated with phone calls and inbound emails from customer service requests that we've been struggling to keep up with," Bryant said. The result was a backlog of "over 2,000 customer service cases that we could not get to within our SLA."
He frames problems like an operations leader, not just a finance one. "I operate my whole life and management philosophy by theory of constraints," he told us — the one or two things keeping you "from growing and getting to the next level." For a few busy seasons running, the constraint was the same: answering customers fast enough to keep them.
Bryant looked at the obvious paths first. Each one had a catch.
Hiring 13-plus people locally in St. Louis would have meant months of job posts, interviews, and onboarding — "busy season's over by then." It also clashed with how Cambridge thinks about people. "Our whole mission is to enrich lives, and so we don't like to hire people seasonally," he said. "When we hire somebody, we're implicitly making a contract with that family." Adding fixed cost for roles that go quiet six months a year wasn't an option for "a penny-pinching CFO."
A traditional BPO had its own friction. Bryant has worked with them before: large setup fees, slower spin-up, and a layer between you and the people doing the work. "You don't get to pick your own team," he said. When issues come up, "you're really dealing with the frontline employee's manager or manager's manager." For a company where customers expect a specific feel on every call, that distance was the dealbreaker. "We didn't want customers calling in and it not feeling like Cambridge, both from a values and a knowledge perspective."
What tipped the decision was control. With Hire Bloom, Cambridge got to interview and choose each person, train them directly, and scale up and down without fixed setup costs. "We got to pick really, really great talent that feel just like any other Cambridge employee," Bryant said. The team of 13 was up "in under two weeks," and the 2,000-case backlog went "down to zero." Then agents started answering calls live. "I haven't seen anything like it before."
The piece he kept returning to was culture. "When you hire somebody for your company, culture's always a lens that you look through. Why, when you're augmenting your staff internationally, would you not use that same lens?" His U.S.-based agents helped interview and train the new team members — a leadership stretch for them, and a reason the new hires felt like Cambridge from day one.
Then we got to the question on everyone's mind. Bryant is no AI skeptic — he uses it daily. "It's right for someone in my seat to be super curious and interested in AI, but not obsessed," he said.
But he's careful about what he optimizes for. "If we deploy AI, it would be wonderfully effective at closing tickets," he said. "But it might not be as effective at turning those customers into raving fans." When you only talk to a customer a handful of times, each call carries weight. "We still need Cambridge human judgment, and our customers still expect human-to-human interaction."
So Cambridge drew a line that's becoming the smart default: use AI "to assist and augment the capabilities of our human operators — not have AI replace the human operators." Our own analysis of nearly 11,000 home-services reviews backs the instinct: 69% of customer comments mentioning AI are negative, clustered in the high-stakes moments Bryant described — emergencies and judgment calls an SOP can't script.
For operators who've never hired internationally, Bryant's advice is about risk, not borders. His framing is one any CFO will recognize. "Don't go through a one-way door and lock yourself into something. Go through a two-way door." A traditional BPO or a full AI build can put you "in for $100,000 before you figure out if it's a good solution." Staff augmentation lets you test the idea and reverse it if it doesn't work. "Step your toe into the water. Don't dive in for a half-a-million-dollar investment with no undo button."
He sees the same constraint everywhere in home services, where companies have 10x the customers and fewer chances to recover a bad interaction. The scarce resource isn't money — it's attention. Operators are "wearing 14 different hats." Remove a headache, free up that attention, and put it back into customers. "Treat those touchpoints, good and bad, as the gold they are."
Bryant covered far more than we can fit here — his theory-of-constraints lens and what his St. Louis team learned working alongside international colleagues. Watch the full Hire Bloom AMA with Bryant Blair here.
If you're staring at your own busy-season backlog, you don't have to choose between speed and the human touch your customers expect. Hire Bloom helps you build a team that feels like yours — interviewed, trained, and screened for your culture by you. Get started with Hire Bloom here.
Tell us where you're at, and we'll show you what's possible.
Find the right talent for your open roles
Get a custom plan with flat, predictable pricing
Start interviewing in as little as 2–3 days
.avif)
