An aviation industry startup accelerator program has selected an artificial intelligence (AI) and natural language processing (NLP)-based platform called Whispr to help improve the speed, accuracy and safety of aircraft inspections and turnarounds.
As part of International Airline Group’s (IAG) Hangar 51 accelerator program, Whispr is partnering with Spanish carrier Iberia to implement its hands-free voice guidance platform on two projects. The first project is being conducted with Iberia Maintenance at Madrid-Barajas Airport (MAD) to digitize the aircraft inspection documentation process on the airline’s new fleet of Airbus A350s, which up until now has been entirely paper-based.
According to Hugh O’Flanagan, Whispr’s co-founder and CEO, the existing inspection process entails engineers walking around the aircraft and checking tens of items in each different section as they manually complete a paper-based report, which then has to be manually input into a system. “We’re streamlining that entire process. It doesn’t make sense to us to have the world’s most efficient aircraft and the world’s most inefficient process,” says O’Flanagan. “We’re working with them to bring that process alive and take the paper out of the equation,” he adds, which should improve inspection speed and reduce human errors involved in completing paperwork.
The second project, being conducted with Iberia’s Airport Services Division, is using Whispr to create a more seamless aircraft turnaround workflow for flight dispatchers. O’Flanagan says Whispr has been working with the airline to build a library that flight dispatchers can access which details each part of the turnaround process for different aircraft types and airlines, helping them to better understand the difference in processes and protocols for changing variables.
Whispr’s voice guidance platform works by leveraging AI and NLP to convert text-to-speech and vice-versa, converting process instructions or checklists into interactive voice guidance workers listen to as they complete a task. “That means workers can be guided through their job, they can ask questions about things they don’t understand and they can give their voice confirmations on tasks they’ve completed, all just using their voice,” explains O’Flanagan.
The platform uses AI to understand frequently asked questions—such as what the tire pressure should be on a specific aircraft type—so workers can simply ask the question while working hands-free and find out the answer in real-time rather than stopping work to look through manuals. The platform currently translates between English and Spanish and O’Flanagan says Whispr is adding additional languages all the time. “The beauty of Whispr is it enables businesses to solve the language, accent and literacy challenges that may exist within your workforce, because now you can deliver the guidance, questions and expertise in the native language of the worker,” he says.
To adapt timing of instructions to match where a worker is within a process, Whispr leverages both existing data on how long particular tasks should take as well as voice confirmations from the worker about what has been completed so far. It can also send reminders about completing a task after a certain period of time and sequences in which tasks need to be completed. Through combining these factors, O’Flanagan says Whispr can pinpoint exactly where a worker is within a process and tap the content, instructions and guidance they may need at that moment.
One other feature of Whispr that could provide efficiencies is its A/B testing functionality. O’Flanagan says Whispr has been in discussions with a large legacy airline group whose A/B testing of different processes took a minimum of five weeks to roll out, which Whispr could reduce to minutes by designing a process and rolling it out the same day.
“Maybe you have two ideas for how best to do an aircraft inspection. Why not A/B test those ideas through our system on a controlled group to understand the data in real-time?” he says. “That’s just not possible with a printed-out piece of paper—you would have to pull people off the floor and into a training session to try to educate them on a new way of doing things. But with our technology in your ear and guiding you through it in the moment, all of these possibilities open up.”
O’Flanagan says he sees many potential use cases for Whispr within airport maintenance and services, but for now the company is focusing on how the platform can be implemented within aircraft turnarounds and inspections to ensure they “nail that for Iberia and deliver them a significant value.” According to O’Flanagan, initial data and interactions from the project—which has been running for approximately six weeks—have been positive and the project is on track.
Iberia and Iberia Maintenance declined to comment on the project, which is scheduled to run through the end of 2019.