Cornel's Plumbing Estimates and a 26.4% Bigger Ticket

BlueBoost, a company that sells an AI estimate writer to plumbing and HVAC shops, published a customer story on Cornel's Plumbing, Heating & Air Conditioning in the Portland metro area. The tool turns what a technician found on the job into a written estimate the homeowner can read. BlueBoost reports that from June 2024 through October 11, 2026, Cornel's technicians created more than 17,500 estimates on jobs totaling $13.4M in sales. The figures come from the vendor's dashboard and are self-reported.

The headline number is the average ticket, meaning the typical dollar amount of one completed job. BlueBoost reports the average ticket on Cornel's jobs written with its tool is 26.4% higher than on jobs written without it. That points at something many shop owners already suspect. A thin estimate on a tablet can stall a big job, and the words a tech leaves behind may matter as much as the diagnosis.

Money left on the tablet

Picture the moment after a tech finds a failing water heater or a tired furnace. The homeowner is standing there, and the estimate on the screen says the part name and a price. Nothing explains why it failed, what happens if it waits, or what the choices are. The homeowner says they want to think about it, and the job goes cold.

That is the problem BlueBoost is built for, and Cornel's numbers are the vendor's case that it helps. The 26.4% figure compares jobs at the same company, with and without the tool, over about 28 months of dashboard data. It is a single average across that whole span, so it does not show whether the lift held steady month to month.

It is also fair to ask which jobs got the tool. Techs may have reached for it more on larger repairs, where a written explanation helps most. If so, part of the gap could reflect the kind of job, not only the writing. BlueBoost's story does not separate those effects, and owners should keep that in mind when reading any with-and-without comparison.

From a tech's notes to a written estimate

An AI estimate writer takes a tech's notes, voice memo or short description and drafts a full estimate in plain language. It explains the problem, the recommended fix and what each option includes. The tech reviews it and sends it, so the person who saw the equipment still owns the price and the promise.

Many shops price work from a flat-rate pricebook, a set list of tasks with fixed prices so the homeowner pays by the job rather than by the hour. An estimate writer does not replace that list. It wraps words around it, so a line item like a pressure valve becomes a short reason the homeowner can follow.

Shops also present options using good-better-best, a format that offers three levels of fix at three prices so the homeowner can choose. The writing tool likely helps most here, since three options with no explanation can look like an upsell. BlueBoost reports that at Cornel's, "jobs written with BlueBoost carry 10% more estimates," which suggests more options reached the homeowner, though the story does not say so directly.

A skeptical crew and a rejected invoice

Halpin Plumbing in Cincinnati is a third-generation shop, and its BlueBoost customer story covers July 2025 through October 11, 2026. BlueBoost reports 778 estimates on jobs totaling $1.2M in sales, and says those jobs averaged 71.4% more revenue than jobs written without the tool. It also estimates 65 hours of writing saved. These are self-reported vendor figures.

The Halpin number is revenue per job, not average ticket, so it is a different measure from the Cornel's figure and should not be lined up against it. The more telling part may be adoption. In BlueBoost's telling, Skylar Halpin described the rollout as natural and easy, and said she does not even know whether there was a learning curve. That is a paraphrase in the vendor's story, not a full quote.

RooterPLUS, a plumbing, septic and drain company around Cumming, Georgia, has a third BlueBoost story. From February 2025 through October 11, 2026, BlueBoost reports more than 12,800 estimates and 3,300 invoices on jobs totaling $11.3M in sales, with about 1,070 hours of writing saved. These are also self-reported vendor figures.

The RooterPLUS story carries a cost owners will recognize. For about ten years, according to the story, the office rewrote short field invoices that commercial clients rejected. A property manager is unlikely to pay an invoice that says pipe replaced and nothing else. Every rewrite was office time spent fixing what the field could have written once.

Twenty-five minutes per estimate at a roofer

The fourth source is outside plumbing and HVAC, and outside this month's news. Ontrakt, which sells estimating software, published a case study on DFC Home Improvement on March 6, 2026. DFC is a Mid-Atlantic roofing and exteriors contractor. It is a vendor blog, and the figures are self-reported.

Ontrakt reports that building, formatting and sending an estimate at DFC fell from 2.5 hours on average to 25 minutes. It also reports DFC's close rate rose from 29% in October 2025 to 44.7% in February 2026. Close rate is the share of estimates that turn into booked jobs. Those are two different months, so seasons and other changes may play a part alongside the tool.

The author of the Ontrakt post also estimates the change is worth about $42,500 to $51,000 in additional monthly revenue from the same lead volume, based on a residential roofing ticket of roughly $8,500. That is the author's own arithmetic on that page, an estimate rather than a reported result. Ontrakt also reports about 60 hours a month of estimating time recovered.

Read your last twenty estimates

None of these numbers tell you what your own shop would see. They do suggest a cheap first check. Pull the last twenty estimates your techs sent on jobs over a set size, and read them as a homeowner would.

Ask three things of each one. Does it explain what failed and why? Does it say what happens if the homeowner waits? Does it offer more than one option, each with a reason? If most of your estimates fail those questions, the writing is a likely place to look before buying anything new.

Then compare the ones that closed with the ones that did not. If the closed jobs tend to have fuller explanations, you have your own small version of the BlueBoost comparison, without a vendor's dashboard. If there is no difference, the stall may be somewhere else, such as price or timing.

Commercial work deserves its own pass. If your office rewrites invoices before property managers will pay them, count how many it rewrote last month and how long each took. That number belongs to you, and it is the cost an estimate writer would have to beat.

Before the tool reaches the tablet

Startup Miracle works with owners on the steps before a tool goes on the tablet. We help a shop read its own estimates, find where jobs stall, and decide whether writing is the problem worth solving this quarter. Then we help set the rules: which techs review AI drafts, and who may send a price without a second look.

That review step matters for any tool that writes to customers. We wrote about approval rails before an AI replies to a one-star review, and the same thinking applies to an estimate. On the money side, our piece on who may approve job cost in BuildOps Finance Hub covers the office end of the same field paperwork.

Estimate writing questions from shop owners

What is an AI estimate writer for plumbing and HVAC?

An AI estimate writer turns a technician's notes into a full, plain-language estimate that explains the problem, the fix and the options. The tech still reviews the draft and sets the price before it reaches the homeowner.

How do I tell if my estimates are costing me jobs?

Read your last twenty estimates on larger jobs as a homeowner would, checking whether each explains the failure, the cost of waiting and the options. Then compare the ones that closed with the ones that did not.

How much can better estimates raise an average ticket?

BlueBoost reports, from its own dashboard, that Cornel's Plumbing saw a 26.4% higher average ticket on jobs written with its tool from June 2024 through October 11, 2026. That is a vendor self-reported comparison at one company, not a promise for any other shop.

If you want the long version of how we think about readiness, Start with a FREE quiz and Get Your AI Score.