Imagine the time contractors and distributors could save if they could snap a few photos of a shingle and, within minutes, have a full report about the exact color, brand, and if it’s still on the market.

That thought drove Austin Gunn on a years-long quest to create Roof Pair, a potentially game-changing technology for the roofing industry. Using artificial intelligence, the app will allow users to obtain shingle information without mailing in samples.

“With a certain amount of set photos from different angles and so on, we’ll be able to identify with a 90% confidence interval of what that product is and what color it is,” Gunn said.

Time is Money

The 25-year-old entrepreneur entered the roofing industry in April 2023 as an outside sales rep for SRS Distribution in the Dallas-Fort Worth area. During that time, customers brought in samples or photos of shingles in the hopes of identifying them. Oftentimes, only those with decades of experience could help track down the right shingle.

“Whenever I didn’t know a lick about a steep-slope product, I was like, ‘Why is there not some type of guide or something that can help me basically onboard faster from a distribution perspective and help me understand exactly what products my customers are using?’” he asked rhetorically.

Material identification services are available in the roofing space to help eliminate guesswork. Services like Name that Shingle and ITEL Laboratories ask customers to send a shingle sample that they then analyze in a lab. Gunn explained that the results are helpful, but waiting for them takes time away from completing a job.

These frustrations followed Gunn after he left SRS to join the contractor side of the business as co-owner of Ultra Roofing.

“In the Texas market, looking for discontinued products on roofs is a big thing for our contractors because they can usually get the roof bought out instead of having to do repairs and so on due to incompatibility or various other reasons,” he said.

Austin Gunn, owner of Roof Pair

Eager to resolve the pain point, Gunn took on the arduous task of collecting documentation about shingles from the nation’s major manufacturers. These documents are essential for preparing reports that ensure contractors comply with the specifications set by local building codes or a general contractor. Moreover, using mismatched or incompatible shingles for a repair job may void warranties and impact the roof’s durability.

With a database in place, Gunn set out to develop a tool that could quickly and accurately match this information with the shingle that needed identification.

“Time is everything in this industry, so I wanted to bridge that gap and put all the information in the contractor’s hands so they don’t spend an extra 20 to 40 minutes driving to their nearest supply house to get an answer,” he said.

Machine Learning

Artificial intelligence model training involves feeding data to algorithms, checking the results and adjusting the model to produce reliable results efficiently. Roof Pair’s AI model has been trained in-house using thousands of images of shingles.

Users are instructed to take seven to eight photos of the shingle from multiple angles. Once uploaded, Roof Pair analyzes the shingles based on color, exposure, dimensions and cellophane strips. Gunn said users can expect reports to be generated within seconds, though it could be up to 10 minutes depending on the number of concurrent users upon launch.

“We are definitely an industry that is slow to change and slow to adoption of technology compared to other industries,” he said. “I think over the next couple of years, there will be a lot more software that has [AI] integrated within it or become a centerpiece for it.”

Boasting a 90% confidence interval, there is still a chance the app might not find the exact match. In that case, the program provides recommendations for the closest matches. In addition, Roof Pair can identify if a shingle line has been discontinued and recommend suitable substitutes.

“We'll be able to nail it down to three options, and definitely the model of it,” he said.

The Future Arrives

Gunn said he and his team expect Roof Pair’s identification services to launch within the next 12 weeks. The app’s database is currently free, but a subscription will be required for the identification service.

“There will be a couple of mid-phase updates, is what we like to call it, before the AI feature releases that just improves user experience,” he said.

Beyond identifying shingles to expedite jobs, he hopes the app will become a source of knowledge, whether enhancing communication between project managers and sales reps about a project or allowing distributors to reference it when questions arise about their products.

To that end, Roof Pair is built with the industry in mind, including its shifting demographics. Gunn ensured the app is available in Spanish and included the Spanish versions of shingle documentation when able.

“We want to make sure that everybody can use this app,” he said.

Roof Pair is currently available on both Android and Apple devices.