An HVAC company cut quoting errors from 34% to 4% and boosted close rates from 28% to 41% with AI-powered estimate automation. Here's how.
Article text
How a 52-Person HVAC Company Stopped Losing 1 in 3 Jobs to Its Own Quoting Mistakes
A homeowner in Tampa called two weeks ago about a full system replacement. Two technicians from the same company had visited within the same week, one for the initial diagnostic and one for a follow-up measurement. The first tech quoted $14,200. The second quoted $17,400.
Same house. Same system. Same company. $3,200 apart.
The homeowner called the office to ask which number was real. The office manager didn't know. She pulled up the quoting spreadsheet, found three different line items priced differently between the two estimates, and told the homeowner she'd call back with a corrected number.
By the time she called back the next morning, the homeowner had signed with a competitor.
The Problem Nobody Was Counting
This HVAC company runs three locations across central Florida, employs 52 people, and fields about 340 service calls per month. They're not small. They're not disorganized. They have a CRM, a dispatching system, and experienced technicians with 8-15 years in the trade.
But their quoting process was a mess.
If your HVAC company is losing jobs to inconsistent estimates or slow quote turnaround, we do free 30-minute discovery calls to pinpoint exactly where the breakdown is happening.
When we started working with them, nobody at the company could tell us their quoting error rate. They didn't track it. They tracked close rate (28%, which they knew was below the industry benchmark of 35-40% ), but they attributed the gap to market conditions and price-sensitive customers.
We asked a different question: how often do two technicians quote different prices for the same scope of work?
The answer, after auditing 120 estimates from the previous quarter: 34% of the time.
Not different by a few dollars. Different by an average of $1,850 on system replacements and $340 on repairs. That's the gap between winning a job and watching it walk out the door.
Where the Errors Were Hiding
We spent four days shadowing technicians on ride-alongs and watching the office staff process quotes. The quoting workflow had 12 steps:
Tech completes diagnostic at the customer's home
Tech handwrites findings on a paper work order
Tech drives back to the office (or texts photos to the office)
Office staff receives the work order
Office staff opens the pricing spreadsheet (an Excel file stored on a shared drive)
Office staff looks up each component, part, and labor rate
Office staff builds the estimate in a Word document
Office staff emails the estimate to the service manager for review
Service manager reviews, sometimes adjusts pricing
Service manager sends approved estimate back to the office
Office staff formats and sends the estimate to the homeowner
Office staff logs the estimate in the CRM
Three things jumped out immediately.
The pricing spreadsheet was a graveyard.
The Excel file had been maintained by four different people over six years. It had 2,300 rows, 47 of which had duplicate part numbers with different prices. Material costs hadn't been updated in 11 months, despite two rounds of supplier price increases. Some line items still referenced a refrigerant the company stopped stocking in 2024.
Technicians were freelancing on pricing.
Senior techs with more experience sometimes quoted from memory instead of checking the spreadsheet. They were often close, but "close" on a $14,000 job means hundreds or thousands of dollars off. Two techs had personal notes on their phones with pricing they'd written down at different points in time. Neither set matched the current spreadsheet.
The review bottleneck added days, not hours.
The service manager reviewed every estimate over $3,000. He was also running three crews. His review queue averaged 2.8 days. By the time an estimate reached a homeowner for a system replacement, competitors who quoted on-site had a 2-3 day head start.
According to HouseCallPro's 2026 HVAC estimating guide , the single biggest revenue leak for most HVAC contractors is the time between completing a diagnostic and getting a proposal into the customer's hands. We were looking at exactly that.
The Diagnosis: One Source of Truth, Zero Enforcement
The root cause wasn't bad technicians or a lazy office. It was structural. The company had no single, enforced source of truth for pricing.
The spreadsheet was supposed to be it, but it had no version control, no update schedule, and no mechanism to prevent techs from quoting their own numbers. The review step existed to catch errors, but it was so slow that it created a worse problem: delayed quotes that lost to faster competitors.
Three connected failures, one root cause: the quoting system was manual end to end, and every manual step introduced either an error or a delay.
What We Built
Build time: 4 weeks. We broke it into two phases.
Phase 1 (weeks 1-2): The pricing engine.
We built a centralized pricing database that replaces the spreadsheet. It pulls material costs from the company's two primary suppliers via API every week, so pricing stays current without anyone manually updating a spreadsheet. Labor rates are calculated based on job complexity, system type, and the company's actual cost-per-hour (including overhead, not just the tech's hourly wage). Equipment pricing follows the manufacturer's current list, with the company's margin applied automatically.
Every estimate pulls from this single source. There's no way for a tech or office staff to override a price without a documented exception approved by the service manager, and exceptions are flagged for quarterly review.
Phase 2 (weeks 3-4): The quoting workflow.
We replaced the 12-step process with a 4-step one:
Tech completes diagnostic and enters findings into a mobile form on-site (structured fields, not handwritten notes)
The system generates the estimate automatically, pulling from the pricing engine, applying the correct labor rates, and formatting the proposal
For jobs under $5,000, the estimate goes directly to the homeowner within minutes. For jobs over $5,000, the service manager gets a push notification and can approve or adjust from his phone in the field
The homeowner receives the estimate by text and email, with a digital acceptance option
The tech never leaves the customer's home to generate a quote. The office staff doesn't manually look up parts or type documents. The service manager reviews on his phone instead of sitting at his desk, and only for higher-value jobs.
The Results at 90 Days
We measured at the 90-day mark against the same metrics from the audit.
Quoting error rate:
34% down to 4% (the remaining 4% were edge cases involving non-standard equipment configurations that required manual pricing)
Average quote turnaround (system replacements):
2.8 days down to 2.4 hours
Average quote turnaround (repairs under $3,000):
1.6 days down to 14 minutes (generated on-site while the tech is still in the home)
Close rate:
28% up to 41%
Revenue impact:
At their average ticket size and volume, the 13-point close rate improvement translated to roughly $163K in additional annual revenue
Office staff time recovered:
16 hours per week previously spent on manual estimate creation, now redirected to follow-up calls on outstanding proposals
Quoting process:
12 steps down to 4
The service manager told us the speed was what surprised him most. "I used to review quotes in a batch at the end of the day. Now I get a notification, glance at it on my phone between jobs, and tap approve. The homeowner has the number before I finish my next service call."
What It Cost
The initial build ran about $18K across the 4 weeks. Monthly running cost is $320, covering the pricing API connections, the quoting platform infrastructure, and the AI processing for estimate generation.
At $163K in additional annual revenue, the payback period was under 6 weeks.
This wasn't a digital transformation or a company-wide overhaul. It was one broken process, fixed with one connected system. The scope was deliberately tight: fix quoting accuracy and speed, measure the revenue impact, then decide what to build next.
The Pattern We Keep Seeing
We've worked with HVAC companies on dispatch , plumbing companies on service agreements , and auto repair shops on parts ordering . The pattern is almost always the same.
The problem isn't that the team is bad at their jobs. The problem is that the process between the work and the customer has too many manual steps, and every manual step is a place where errors sneak in or delays stack up.
Quoting is one of the highest-leverage processes in any service business. It's the moment between "the customer wants to buy" and "the customer decides to buy from you." Anything that slows that moment down or introduces doubt (like two different prices for the same job) is directly costing you revenue.
If your quoting process involves a spreadsheet that hasn't been updated recently, technicians who quote from memory, or a review step that adds days instead of minutes, the math on fixing it is straightforward.
What to Do Next
Start with a simple test. Pull your last 50 estimates and check: how many times did two people quote different prices for comparable work? If the answer is more than 5%, you have a quoting consistency problem. Then time the gap between "tech finishes the diagnostic" and "homeowner receives the estimate." If it's more than a few hours, you have a speed problem.
Both are fixable. Often faster and cheaper than you'd expect.
Send us a message if the numbers surprise you. We'll walk you through what a fix looks like for your specific setup, and we'll be honest about whether AI automation is the right tool or if a simpler process change gets you there.