SkyeBrowse CEO Warns of Scaling Bottlenecks in Drone

SkyeBrowse CEO Bobby Ouyang warns that a drone program's greatest challenge often comes after it proves its initial value. In an exclusive interview with DRONELIFE, Ouyang said organizations struggle to move sophisticated capabilities beyond a small group of experts.

For many, launching a program is no longer the hardest part. A police department might start with a few trained pilots. A construction company may rely on one mapping expert. This works initially. Then demand grows. More departments want access. More sites need documentation. Technology that worked for experts becomes a bottleneck.

When Expertise Becomes the Bottleneck

Ouyang says dependence on specialists is a hidden barrier. This applies to crash scenes, construction sites, and facility inspections. He has spoken to teams where one person handles all drone work. When that person takes vacation, the program stops.

The problem is not necessarily the technology. A laser scanner or photogrammetry pipeline can be accurate but still a bad fit if it only creates a specialist role. A small group can learn complicated settings. At 100 users, an organization may have multiple locations and varying experience levels.

Some people use the software daily. Others use it once a month and forget everything in between. Every confusing step becomes a support ticket or an inconsistent result. This creates a paradox. As drone technology becomes more capable, the user experience may need to become simpler.

"As the technology gets more capable, more of the complexity should move into the software and away from the user," Ouyang said. Early adopters enjoyed complexity. The next thousand users are patrol officers and site superintendents who have a real job. They might spend only about ten minutes on 3D mapping.

For SkyeBrowse, this philosophy means pushing technical work behind the interface. Its workflow uses video capture and cloud processing. The goal is for the user to do less. Ouyang described it as a two-orbit automated flight, a video upload, and a model a few minutes later. Sophisticated on the inside, simple on the outside.

Look Beyond the Feature List

Ouyang argues organizations should evaluate more than just features, accuracy, and price. They must consider the human and operational burden required to use those capabilities.

"A feature you cannot staff is a feature you do not have, and the license is only one part of the cost," he said. Other costs include onboarding, refresher training, troubleshooting, and administration. There is also the time people spend waiting for results.

He recommends asking key questions. How long does a new user need to produce a usable result independently? Can an occasional user complete the workflow a month after training? How much manual work happens between capture and a usable result? One question is especially revealing: can the process continue if the organization's most experienced person leaves?

"Then test the complete workflow with your own people, including whoever receives the information, at the worst time of day you will actually run it," Ouyang advised. Enterprise buyers ask these questions about other software. Drone software should get the same treatment.

Capture Is Only the Beginning

Ease of capture alone does not make a program scalable. Ouyang estimates capture is maybe a third of the workflow. The person flying is just the first person in the chain.

A project manager may need to compare conditions across multiple visits. An adjuster may need to assess damage. An investigator may need to understand a scene they never visited. These users should not need to become mapping experts just to access information.

"If any of those people need software installed, a license, or a training session to look at the model, it dies on the operator's laptop," Ouyang said. SkyeBrowse allows models to be viewed through a browser link. Its Public Safety and Enterprise offering includes unlimited users. The idea is that captured data becomes more useful when it can move beyond the original operator.

The return on a capture grows every time another team can use it without asking the operator.

When Success Creates More Demand

The Middletown Police Department provides an example. According to figures provided by SkyeBrowse, the department documented crash scenes in 2023 using a laser scanner and five officers. Each scene took about two and a half hours. Across 65 crashes, that represented more than 812 officer hours.

After adopting SkyeBrowse, the department reported one officer could clear the same type of scene in about seven and a half minutes. That officer does not need to be the department's drone expert.

This distinction is key as success drives more demand. One team gets a great result, and leadership sees it. Suddenly everyone wants in. Construction wants progress documentation. Facilities wants building walkthroughs. Risk management wants a record of site conditions. In an agency, every crash and every callout 'needs the 3D.'

If all these requests depend on the same small group of experts, demand turns into a backlog. The program gets a reputation for being slow.

Ouyang recommends organizations prepare before expansion. Train more users than initially seem necessary. Establish standard capture procedures. Consider if workflows depend on a single aircraft. Establish data governance before the first important mission. Involve the people who will eventually use the information.

If a facilities manager needs specific equipment documented, that should shape the capture. Then measure the time from capture to actual use. That number shows where expansion will create work before rolling it out to more teams.

This is increasingly important as Drone as First Responder programs and routine operations drive higher flight volumes. Flights per organization go up by orders of magnitude. Nobody is going to hire specialists at that rate. The principle applies beyond public safety to construction, insurance, and utilities.

Technology that becomes a bottleneck has a person, a workstation, or a manual step in the chain. That bottleneck scales up quickly. The real measure of scalability is not simply what the technology can do. It is how much additional expertise and human effort each new mission requires.