Azaz Vepari06 Mins read
XBuild Says It Can Build a Roofing Estimate in 15 Minutes
XBuild raised $19M and launched Roofing Proposals, which turns measurement PDFs into a signed, paid proposal. What the release actually promises, what a fast draft still cannot know about a roof, and the 20-job test that settles it.
In this article11 sections
XBuild has raised $19 million and launched a product that says it can turn your measurement report into a customer-ready roofing proposal in under fifteen minutes.
The interesting question is not whether an AI can type an estimate. Anything can type an estimate. The question is whether this removes enough slow, repetitive work to help you close more jobs — without introducing the one kind of mistake that costs real money.
What XBuild actually announced
| Item | Detail |
|---|---|
| Announced | 20 January 2026 |
| Raise | $19 million Series A |
| Led by | N47 |
| Participating | Rackhouse Ventures, Andreessen Horowitz |
| Product | Roofing Proposals, for residential roofers |
| Claim | “Generate a finished estimate in under 15 minutes” |
| Founded | 2022 |
That is a serious round for a company attacking one document in one trade. It is worth understanding why the document is the target.
The workflow is bigger than the sentence it writes
This is the part that separates XBuild from asking a chatbot to draft a proposal.
Per XBuild’s release, the product “references PDF reports including EagleView, GAF QuickScope, Hover, and RoofR and supports manual photos including via CompanyCam.” Worth reading precisely: the first four are measurement reports it parses; CompanyCam is how site photos get in. They are not the same input doing the same job.
From there it is “integrated with the industry’s largest material suppliers such as ABC Supply to provide real-time supplier pricing” — so the material line items are priced against a live catalogue rather than a spreadsheet somebody last updated in spring.
And then it keeps going past the estimate. XBuild says the product “generates a homeowner link with Good/Better/Best pricing options, e-sign, and Stripe Connect payments.”
That chain is the actual product. Measurement, scope, material pricing, presentation, signature, payment — assembled once instead of copied between five systems. The paragraph of sales copy at the end is the least valuable thing in it.
The numbers, and whose numbers they are
XBuild states that over 15,000 projects have run to completion on the platform, representing $250 million worth of construction projects, and that contractors have saved more than 40,000 hours creating estimates.
Those are XBuild’s own figures, published in its funding announcement. They are not an independent study, and they describe the broader platform rather than the new roofing product specifically. Treat them as evidence that real contractors are running real volume through it — not as a forecast for your shop.
I also could not find a published price for Roofing Proposals. You will need the full subscription cost and the transaction fees directly from the company before any of this becomes an ROI calculation.
Why fifteen minutes is genuinely worth something
Take the claim seriously for a moment, because the alternative it replaces is genuinely bad.
The normal path is: the estimator finishes the inspection, drives back to the office, opens the measurement report, opens the price list, rebuilds a material list by hand, assembles a proposal, and sends it tomorrow. Meanwhile the homeowner has three other roofers coming.
Speed to the homeowner while the job is still live is a real commercial advantage, and it is the correct thing to attack. Our roofing estimating software comparison covers the tools competing for the same fifteen minutes.
Why a fast draft is not a correct estimate
Here is the part that decides whether this makes you money or costs it.
A measurement report tells you geometry. It does not tell you:
- the condition of the decking underneath
- ventilation that is inadequate and has to be corrected
- flashing details the report never saw
- local code requirements and permit costs
- access — can a truck get to it, where does the dumpster go
- disposal, and how many layers are coming off
- steep charges, crew productivity, hidden conditions
- whether this turns into an insurance supplement
None of that is a criticism of the AI. It is a statement about the inputs. The contractor is still responsible when a missed line item eats the margin, and a measurement PDF has never been able to see a rotten deck.
So the honest framing is not “AI estimator.” It is an estimating assistant that handles document extraction, product selection, repetitive line items and proposal assembly — while an experienced estimator inspects the inputs, verifies quantities and waste factor, adjusts labour, confirms scope and approves the final number.
The real risk is a believable number
This deserves saying plainly, because it is the failure mode nobody plans for.
The costliest AI error in the trades is not a strange sentence. It is a plausible number that nobody checks.
A messy spreadsheet invites scrutiny. A polished, branded, good-better-best proposal does not — it looks finished, so people treat it as finished. A professional presentation can hide a bad assumption just as easily as it can present a good one, and the better the document looks, the less likely anyone is to interrogate the figure inside it.
That is an argument for a review step, not against the product. But the review step has to be real, not a formality that happens after the proposal has already been sent.
The wider play
XBuild says it intends to expand beyond roofing into concrete, landscaping, painting, windows and doors, glass and glazing, insulation, HVAC and plumbing.
Every one of those trades has the same underlying problem: the information needed to price a job arrives as PDFs, photos, emails, price lists, scribbled notes and human judgment. Whoever can reliably turn that mess into a defensible scope owns a valuable position in the contractor’s workflow — closer to the money than a CRM, and harder to replace.
That is a bigger ambition than roofing proposals, and it is probably what the $19 million is actually funding.
The 20-job test
Do not evaluate this on a demo. Rebuild history, where you already know the answer.
- Take 20 completed jobs with known final costs and rebuild the estimates through XBuild.
- Measure four things: time saved, line items missed, price accuracy, and gross margin variance against what the job actually cost.
- Count how much editing the estimator still performs. That number is the real automation rate.
- Only then run new live jobs, without changing your approval process.
And ask the questions that do not appear in a demo:
- How current is the supplier pricing, and how often does it refresh?
- What happens when a measurement PDF is incomplete or unreadable?
- Can every AI-generated line item be traced back to its source document?
- Who owns the estimate data, and how are homeowner details and photos stored?
- Can you export your work if you leave the platform?
That last one matters more than it looks. A system of record you cannot leave is a pricing decision you make once and live with forever.
Frequently asked questions
How much does XBuild cost?
XBuild does not publish a price for Roofing Proposals in its funding announcement or the pages reviewed. You need both the subscription cost and the Stripe Connect transaction fees from the company directly before calculating a return.
Which measurement reports does XBuild read?
Per its release, it references PDF reports including EagleView, GAF QuickScope, Hover and RoofR, and supports manual photos including via CompanyCam. Note that GAF QuickScope is a different product from GAF QuickMeasure.
Can XBuild really produce an estimate in 15 minutes?
“Under 15 minutes” is XBuild’s own wording for generating a finished estimate. Whether that estimate is correct for your jobs depends on your price book, your waste factors and your review discipline — speed and accuracy are separate questions.
Does XBuild replace a roofing estimator?
No, and treating it as though it does is where the money gets lost. It removes document extraction, product lookup and proposal assembly. It cannot see decking condition, ventilation, access, disposal or hidden conditions, so the final number still needs a human who understands roofs.
Is XBuild only for roofing?
Today, yes — Roofing Proposals is built for residential roofers. XBuild says it plans to expand into concrete, landscaping, painting, windows and doors, glass and glazing, insulation, HVAC and plumbing.
Are XBuild’s published results independently verified?
No. The 15,000 projects, $250 million in construction value and 40,000 hours saved are XBuild’s own company figures from its funding announcement, and they describe the broader platform rather than the new roofing product alone.
The verdict
XBuild is attacking a real bottleneck, and attacking it at the right place — the gap between the measurement report and a proposal the homeowner can sign.
A fifteen-minute proposal is a genuine advantage for a roofer who already has disciplined pricing and a real review step. It will not rescue a company with bad cost data; it will produce bad numbers faster and dress them better.
The best version of this product makes your estimator faster. It does not make judgment optional.
Sources
Sponsored
Ready to upgrade your dispatch?
FieldCamp uses AI to optimize routes, auto-assign jobs, and eliminate scheduling chaos — so you can run more jobs with fewer headaches.
Tested and written by
Azaz Vepari
Editor, Modern Field Service
Listen to this review
The same comparison as a podcast episode, for the drive between jobs.
Tool profiles
Pricing, best fit and limitations for each tool in this review, one page each.
