Updated October 2026. XactGen Roofing is in rollout; details below reflect Verisk's public materials as of publication.
For twenty years the roofing estimate inside Xactimate has been built the same way by everyone: sketch the roof, then enter the line items one by one. Verisk is now automating the second half.
XactGen Roofing is a new capability built inside Xactimate that generates roofing line items and quantities from roof measurement data. Verisk introduced it in a June 2026 webinar titled "Roofing Estimates, Reimagined," and its site lists it as coming soon to Xactimate. The headline promise is blunt: get roofing estimates 80% complete in minutes.
If you write or negotiate insurance roofing claims, this is the most important Xactimate development since the move to the cloud. Here's what it is, who it's for, and what to do about it.
What XactGen Roofing Actually Does
Strip away the marketing and the workflow is simple:
- Start from Sketch. With roof measurement data in the estimate (an aerial report import, or data acquired through XactGen itself), the user launches XactGen Roofing from the Sketch workspace.
- Generate. XactGen reads the measurement data and the estimate's context — roof type, materials, the claim's parameters — and produces roofing line items with quantities.
- Review, edit, finalize. The results are laid out for a human to check, adjust, and approve. Verisk is emphatic that the estimator stays in control.
In other words: XactGen does what a good macro library and a disciplined estimator do together — turn geometry into a complete scope — and it does it automatically, inside the software, with Verisk's own line-item logic behind it.
It sits in a broader AI push from Verisk this year. XactAI added photo labeling, assignment summaries, and plain-language prompts across the Xactimate suite. In May, Verisk connected XactAI-powered estimating to Anthropic's Claude, letting contractors describe a job in natural language and turn it into structured estimate actions. And the ecosystem keeps tightening on the measurement side: Roofr added one-click Xactimate ESX file ordering in April, and ServiceTitan's Xactimate integration was slated for summer. The pipes from "roof measured" to "estimate written" are being welded shut, with AI in the middle.
Who It's Built For (Read This Part Twice)
Here's the detail that matters most for roofers. Verisk's own framing of XactGen Roofing is about claims teams: closing roof damage claims faster, keeping "the adjuster in control," reducing one of the biggest bottlenecks in property claims. Verisk notes that roofing accounts for roughly 30% of all line items in claims estimates — which is exactly why carriers want it automated.
So the first and heaviest users of XactGen will be the desk adjusters and carrier estimating teams writing the initial scope on your claims. The practical consequence:
The carrier's estimate you receive is increasingly going to be AI-generated, from the same measurement data you're looking at, and it's going to be "80% complete."
That last phrase is the whole ballgame. Eighty percent complete is a Verisk number, not ours. Which means the remaining 20% — the judgment calls, the site conditions, the items the model didn't have context for — is where every roofing claim will be won or lost from here on.
Three Things Roofers Should Expect
1. Faster, more consistent initial scopes
The good news first. AI-generated scopes built from real measurement data should mean fewer wildly wrong square counts, fewer missing-facet estimates, and faster first offers. Verisk's standardized logic also means the carrier estimate will be internally consistent in ways human-written scopes often aren't — the same roof should produce the same baseline from the same adjuster's desk every time.
2. The same 20% missing, more predictably
What a measurement-driven generator can see: squares, pitch, facets, ridges, hips, valleys, eaves, rakes. What it can't see from a measurement file: the second layer under the shingles. The rotted decking at the northeast corner. The rusted step flashing. The drip edge that isn't there. Whether the chimney needs a cricket. Whether the attic's ventilation meets current code on a tear-off.
Those are site-condition and code-compliance items — the components homeowners never see and the ones that already go missing from carrier scopes. An AI that generates from geometry will be excellent at geometry and structurally blind to everything that requires a ladder and a camera. Expect the supplement conversation to shift toward documentation and away from arguing about square counts.
3. Your documentation becomes the whole argument
If the carrier's scope is generated from the same aerial report you ordered, disputes about the roof's dimensions are over before they start — that's a win. The entire negotiation moves to evidence of what the model couldn't know: photos of the decking, the layer count, the flashing condition, the code citation for the drip edge. Roofers who photograph like inspectors and cite code like consultants will thrive. Roofers whose supplements lean on "it's a cut-up roof, trust me" are about to have a rough year.
Why Your Own System Matters More Now
It would be easy to read all this as "Verisk automated the estimate, so stop building macros." That gets it exactly backwards, for three reasons:
XactGen generates their baseline, not your scope. A roofer's estimate still needs to be complete, defensible, and yours — written to your inspection, your market, your standards. A generator built for claims teams produces the carrier's starting point. Your macro library is your complete scope — including every item the generator has no way to see — and the gap between the two is the supplement.
Reviewing fast requires knowing the codes cold. When the carrier's scope arrives 80% complete, your job is to find the 20% in minutes, not hours. That means reading an estimate at speed: recognizing instantly that the starter line is missing, that the ice & water is in squares instead of square feet, that the steep charge is absent on a 9/12. Fluency in the core line items — and keeping the reference where you can see it — becomes a review skill, not just a writing skill.
Your knowledge is the asset no platform ships. Verisk's generator will improve, and your access to it will depend on your subscription tier, your carrier's workflow, and rollout timing you don't control. The estimating knowledge built into your own macros, your inspection checklist, and your documentation habits travels with you through every version of this software and every ownership fight in the ecosystem.
What to Do This Quarter
- Watch for XactGen in your Xactimate and try it on a closed claim. See what it generates for a roof you know — and note exactly what it misses. That list is your supplement checklist for the next five years.
- Audit your macros against the misses. Make sure every item a measurement-driven generator can't see — layers, decking, flashing condition, ventilation, code items — lives in your complete roofing macros by default.
- Upgrade your field documentation. If the argument is moving to evidence, standardize the photo set: every elevation, every penetration, every flashing detail, the eave edge, the attic.
- Get fast at reading estimates. Review speed is the new estimating speed. The shortcuts and the reference under your keyboard matter as much for the carrier's file as for your own.
The Bottom Line
XactGen Roofing isn't a threat to roofers who estimate well. It's a threat to roofers who estimate vaguely — because the carrier's side is about to get precise, consistent, and fast on everything a measurement file can show. The 20% that requires a ladder, a camera, and a code book is the roofer's territory, and it always was. The software just drew the border line more clearly.
Build your system for the 20%. That's where the money was anyway.
Share:
Xactimate Certification Levels 1, 2, and 3: Are They Worth It for Roofers?