Experts are convinced that self-flying planes will roam the skies by 2025 - here's how one startup is working to win over the FAA and the public

• Xwing, founded by Marc Piette, is one of the startups working to make self-flying planes a reality.
• Its Cessna 208B Grand Caravan can already fly on its own, as Insider found on a demonstration flight.
• Self-flying planes will start by flying cargo and then regional passenger flights as early as 2025.

Teaching a 27-year-old aircraft how to fly on its own was the easy part for Marc Piette and his team at Xwing. The real challenge is how to get the technology flying on commercial aircraft, and accepted by the public.

Piette had the idea to conquer self-flying aircraft when driving from San Francisco to Eureka, California, a near-300 mile journey that takes five hours by car. As a student pilot taking flight lessons at Palo Alto Airport at the time, he couldn't accept that driving was the most efficient way to travel regional distances for the average person.

"The time it takes me to get to places like [Eureka] from San Francisco is about the same time it takes me to get to New York," Piette, the founder and CEO of Xwing, told Insider. "It's absurd. Traveling 250 miles shouldn't take me the same amount of time it takes to travel across the cross country."

And the idea for Xwing was born. The vision was to use the vast aviation infrastructure that already exists but make it more accessible and bring costs down by using autonomous technology.

"The only way to travel fast on the ground is through massive infrastructure investment ... which isn't happening anytime soon," Piette said.

Learning to fly

Xwing has already taught a Cessna 208B Grand Caravan how to operate entirely on its own, as Insider saw firsthand on a recent demonstration flight. After the engine was started and the autopilot engaged, the tech took over and no pilot input was required.

Getting to this point, Piette explained, was easy. The few challenges that remain come with certifying aircraft with regulators and ultimately getting the public to accept self-flying aircraft.

"We don't see very many hurdles left, frankly," Piette said. "There are details on implementation, details about like certification systems [and] specific sensors, but they're not tough problems."

Air traffic control is also one of the biggest roadblocks in the way of truly autonomous flight. The current system in the US is based on voice communications between controllers and pilots.

Xwing's self-flying aircraft, however, can't talk at all. The temporary solution is using a ground-based operator that can communicate with controllers and relay commands to aircraft. The initial launch of Xwing's tech will likely see one operator on the ground control multiple aircraft in the sky.

It's not a perfect solution, as the goal is for complete autonomy, but it will still significantly reduce aircraft operating costs. Airlines and aircraft operators can potentially save millions on salaries and benefits by reducing their pilot pools.

Lessons learned from the Boeing 737 Max

Piette doesn't view the Boeing 737 Max crashes as a setback to his goal of autonomous flight. Rather, he says the response to the crashes a "testament" to how safe commercial aviation has become since those fatal incidents are so rare.

"In the end, people understand that despite these accidents, [aviation is] still the safest way to travel," Piette said. "And that's due to decades of diligent work in systems engineering, system safety analysis, [and] accident reviews to keep improving the safety of aviation."

Piette pointed to Tesla and the rare incidents that resulted from the Autopilot functionality of its electric vehicles. And yet, the cars are still among the most popular on the road.

Increased automation helps reduce pilot error, which Piette says is the cause for a majority of aviation incidents over the years.

"Part of the reason why the rate of accidents has declined so significantly over the past few decades is due to increased levels of automation in these aircraft," Piette said.

Read More: 6 autonomous aviation startups poised to bring self-flying planes, drones, and air taxis from the fringes to reality, according to VCs

But the Max did have an impact on how new aircraft are certified. The FAA is strengthening its "designated engineering representative" program with more scrutiny to ensure the aircraft that are being developed are safe to fly in.

"That can slow things down a bit, but in the grand scheme, it doesn't change things," Piette said.

Flying packages first instead of people

Small cargo planes will likely be the first aircraft to see pilotless operations as having an inanimate payload means there's less risk involved.

"Cargo is, to some extent, an easier first place to deploy this because you can get exemptions [and] work with the regulator to start flying initially over unpopulated areas," Piette said. "That way you're not just putting anyone on board in jeopardy, but you're also not putting anyone on the ground at risk."

Xwing cites short-haul cargo runs to remote communities as the first example of what it hopes to replace.

Mountain Air Cargo, for example, flies a cargo route between Raleigh and Wilmington in North Carolina. A morning flight from Raleigh takes the packages that arrived earlier on larger FedEx Express planes and delivers them to Wilmington.

Then, a pilot waits around all day until it's time to fly back, after the cutoff time to send packages in the evening. Both flights take an average of 30 minutes and can be easily replaced by autonomous aircraft.

Xwing's technology, regardless of whether it will fly cargo or passenger planes, will aim to meet the higher standards required for passenger planes. It won't be too long, Piette estimates, before self-flying planes are flying people once the tech is established in the cargo-flying realm.

"Passenger applications for this will come fairly soon after the commercial deployment of cargo applications," Piette said. "If I were to predict, within two years of commercial deployment of this application on the cargo side, you'll start seeing passenger applications, as well."

Slowly increasing automation in the cockpit

Regional passenger aircraft and business jets will likely be the first planes to move into the self-flying realm, according to Piette. And the first step is to move away from having two pilots in the cockpit.

Business jet manufacturers have primarily taken the lead on incorporating more automation into their cockpits. The Cirrus Vision Jet or Cessna Citation aircraft, for example, are intended for single-pilot flying and operate alongside their larger counterparts in the upper altitudes.

Dassault Aviation is leaning more into automation with its new flagship, the Falcon 10X business jet, which is designed to accommodate single-pilot operations during cruise flight. "The level of automation will allow us to have one pilot flying the aircraft, while the other is resting in the cockpit," Philippe Duchateau, Dassault's chief test pilot, said during the aircraft's unveiling.

While the shift to fully self-flying systems won't take place overnight, the foundation is already there with current autopilot and single-pilot cockpit systems. And inch by inch, the industry will move closer to fully self-flying aircraft.

"By the middle of this decade, you're definitely going to have pilotless cargo aircraft, and very shortly after that pilotless passenger aircraft on the regional side," Piette said.