Where a Boiler Repair Business Actually Loses Money

Before comparing costs to benefits, you need to know what the leak looks like. For most small trades businesses, the biggest losses aren't technical, they're operational:

• Missed calls. A customer with no heat calls three shops. The first one that answers gets the job. Every unanswered call after hours or during a job is revenue walking to a competitor.
• Manual scheduling and dispatch. Hours per week spent on phone tag, confirmations, and rebooking, time that isn't billable.
• Slow follow-up. Quotes sent late, reminders forgotten, and maintenance-plan renewals that never get a second touch.

These are the workflows where AI has a realistic payback story, because the cost of the status quo is measurable. AI that replaces a technician's hands doesn't exist yet, AI that answers the phone at 9 p.m. does.

The Cost Side: What You're Actually Paying For

A realistic cost-benefit analysis for a small shop includes four buckets:

1. Setup and build. Whether it's an off-the-shelf chatbot or a custom build, there's an upfront cost in money and in your time defining how it should behave.
2. Running costs. Monthly fees, phone or software integrations, and occasional updates as your services and pricing change.
3. Training and process change. Staff need to know how to hand off to and from the AI. This is usually the most underestimated line item.
4. Quality control. Someone has to review transcripts, catch mistakes, and approve anything the AI sends to customers. Budget for it.

Payment models vary. Some vendors lock you into long contracts; others let you start monthly, build in milestones, or own the system outright. For a small business, the models that let you prove ROI on one workflow before expanding are the ones that protect your downside.

The Benefit Side: Realistic Labor Savings

Here's the honest math. A full-time receptionist or dispatcher in Canada typically costs tens of thousands of dollars per year once you include wages, benefits, and overhead. An AI receptionist or dispatcher, one that answers every call, books jobs, and routes emergencies, typically runs at a small fraction of that. That's the headline saving, but the second-order ones matter too:

• Call capture. If the AI converts even a few after-hours calls per month that you'd otherwise miss, it can pay for itself.
• Admin hours reclaimed. Even 5–10 hours a week of scheduling time back is real capacity, either billed out or taken off your back.
• Consistency. The AI doesn't forget to follow up on a quote or leave a customer on hold during a rush.

Be skeptical of any vendor quoting you a precise ROI number before they've seen your call volume. A credible partner will model it from your actual numbers, call logs, job values, close rates, not industry averages.

Payback Period: How to Think About It

For a focused first project, say, an AI receptionist capturing missed calls, the payback question is simple: does the value of recovered calls plus reclaimed admin hours exceed the monthly cost within a few months? For most shops with steady call volume, the math is favorable. For a shop with five calls a week, it may not be, and a good vendor will tell you that.

Where payback gets murky is broad "AI transformation" projects with no defined workflow. That's where the industry-wide finding that most broad AI adopters report no revenue lift comes from, spend without scope. The fix isn't less AI; it's narrower AI.

The Governance Gap

AI agents fail in production for predictable reasons: they're given too broad a mandate, nobody reviews their output, or they're connected to systems they shouldn't touch. Before any AI touches your customers:

• Scope it to one job, answering calls, or booking, not both plus invoicing on day one.
• Review transcripts weekly for the first month.
• Keep a human escalation path for anything the AI can't handle, especially emergency no-heat calls.