How AI is Rewriting the Rules of Documentation

You might still remember the yellow binder days—when documenting a claim meant seven layers of carbon copies and a handshake at the bar. Mitch Byrom also remembers those days. He joked that, back then, “contractors knew how to do construction and restoration, and adjusters knew how to assess damage.”

But as Byrom, founder of Restoration-OS and a longtime IICRC instructor, explained, today’s water damage claims are a very different story—and artificial intelligence (AI) is right in the middle of it.

AI is no longer some looming trend or gimmick—it’s here to stay, embedded in your smartphone, calculating your estimates, drafting your sketch, reviewing scopes, and shaving days off your billing cycle.

So what does that mean for you, the person actually wading through water, documenting drying logs, and arguing line items?

With decades of restoration experience and multiple software, businesses, and executive relationships, paired with his knowledge of claims, Byrom shared his expertise on the subject with the restoration community.

From floppy disks to LiDAR scans

You didn’t ask for this shift. “We were forced to use Xactimate,” Byrom reminded. “Not because people didn’t like your handwritten notes, but because the carriers wanted a standardized price list.”

Xactimate, among other products available in the marketplace, is one of the software systems that insurance adjusters, contractors, and restoration professionals use to create detailed property damage estimates. It uses advanced algorithms and a comprehensive database of repair and replacement costs. The process is highly automated, reducing the potential for errors and ensuring a high degree of accuracy. Its use was the first major shift to technology in the industry.

The earliest version was built off a group of contractors inputting their pricing on floppy disks. That evolved into national pricing averages, labor efficiency tracking, formula research, and the rise of unit pricing.

But what about AI? “AI is different,” Byrom said. “AI is when the system learns. When it starts making decisions based on data, it’s not just your calculator—it’s your assistant, estimator, and auditor. And it’s learning fast.”

Take measurement, for instance. “If you get a bad measurement on ceiling height, you miss out on cubic footage for dehumidification, on antimicrobial, on cleaning labor, on paint. You even lose your scaffold charge,” Byrom explained. “All because you measured 10 feet instead of 12 feet.”

Now, your iPhone can do it better. If you have a 12 Pro or above, you already carry a LiDAR (light detection and ranging) scanner in your pocket.

“Your phone doesn’t forget a measurement,” he said. “It doesn’t misread the tape. And it’s fast.”

Meet the new technician: AI

Byrom doesn’t just teach this stuff—he uses it. His 21-year-old son, Kyler, was recently conducting an inspection at a Hardee’s in flood-soaked Western Kentucky. Kyler’s no Applied Structural Drying (ASD) expert, just a normal, barely-passed Water Restoration Technician (WRT). But with Restoration-OS in hand, a LiDAR-equipped phone, and AI-powered sketch tools, he can generate a complete carrier compliant water sketch, scope, and moisture map documentation set on his own.

“It’s not because he’s the world’s best tech,” Byrom admitted. “It’s because he had the right tool.”

That’s the new reality: High-skill outputs from low-skill labor—when paired with the right tech. You don’t need a black belt in restoration to deliver a perfect sketch and inspection anymore. AI is doing the heavy lifting.

Speed, accuracy, and the estimate gap

At some point in your business, you’ve likely been told that claims need to be processed faster; that carriers expect a three-day turnaround on estimates. And if you’re like most contractors, you’re actually averaging nine.

“They say we need to speed up,” Byrom said. “But really, what they mean is this: We will be mandated to speed us up, if we don’t start doing it fast enough.”

Carriers and private equity groups are already investing heavily in AI for estimate generation and review technology. Byrom named tools like Yembo and Hosta A.I., which scan rooms, calculate costs, and even initiate homeowner payouts—all before you get on site. Some of these are coming out of academic leaders like the Massachusetts Institute of Technology.

“That check’s going straight into their account,” he said. “And then what? The homeowner calls around, looks for someone cheaper, and now you’re stuck supplementing.” Does that sound familiar? The auto body industry has seen how this changes up the business process since they have already seen this shift in auto claims.

How the bots really think

If you think AI is always accurate, think again. “Garbage in, garbage out,” Byrom said. “AI is only as smart as the data it gets.”

So, double-check everything, especially if you’re relying on tech to build your sketch or pick your line items.

“Is it fiberglass reinforced plastic or drywall?” Byrom asked. “Is it Category 3 or Category 2? Was it after hours? Weighted extraction? The AI won’t know unless you tell it.”

Even when an AI system recommends line items, you’re still the quality control. “You’re not the estimator anymore,” Byrom explained. “You’re the auditor.”

Context is king

Byrom shared a story about working with RICOH, the global tech company behind some of the best 360-degree cameras on the market.

He convinced them to integrate their imaging into his app, and now they’re exploring ways to teach AI how to recognize furniture by its cubic volume, weight, and even SKU (stock-keeping unit) numbers.

“Imagine scanning a room and instantly knowing how many vaults you’ll need or getting a suggested contents list and estimate built automatically,” he said. “That’s where this is going.”

Still, AI won’t know if the sofa’s medium-density fiberboard (MDF) is swollen from water damage. It won’t catch the nuance of customer distress. It won’t interpret context as well as you can.

“AI won’t get you paid faster,” Byrom clarified, “but it can help you submit faster. And submission speed is directly tied to payment speed.”

Who’s training whom?

Byrom provided one final takeaway: The best AI won’t come from big tech or software vendors—it’ll come from you.

“The algorithm only gets smarter when you feed it,” he explained. “If you build your own workflows, train your own models, and set up your own logic, it works for you—not against you.”

His app, Restoration-OS, is designed with that in mind: choose the materials, conditions, and damage levels, and the app automatically fills recommended equipment and scope line items. You’re not just guessing—you’re guiding.

And when you’re guiding, you’re still in charge.

The future is fast—and human

Yes, AI is here. It’s scanning rooms. It’s building estimates. It’s replacing pencil sketches with phone scans and turning technicians into data collectors. But as Byrom reminded us, “AI doesn’t replace judgment. It replaces repetition.”

The real challenge isn’t learning how to use it. It’s knowing when not to.

So next time someone says you’re behind the curve, remember what Byrom said: “You don’t have to be afraid of AI. You just have to be better than it.”

Watch the full interview below: