Field Notes: Skywise in Sao Paulo

Taking part in an airline’s digital transformation can be an exhilarating, challenging, and deeply gratifying experience. My time in Sao Paulo with Azul Brazilian Airlines was as memorable as it was productive. The Azul team brought enthusiasm and a willingness to challenge their own status quo by using the power of data to fundamentally evolve operations. I left Brazil confident that Skywise will empower the Azul team to remain at the forefront of digital innovation, continuing its journey as one of the world’s most analytically advanced airlines. I’m excited to share some of what we accomplished during Azul’s onboarding onto the Skywise platform, and some of the ingredients that made it so successful.


But first, what is an onboarding? After the ink has dried and an airline has made the decision to join the Skywise community, a team of engineers and project managers arrive onsite equipped with the Skywise platform. Our team has three goals in mind for every airline: integrate the data, design workflows, and empower users.

As Lead Data Scientist, my aim is to make sure that we hit each of these goals in a way that maximizes operational value for the airline. First, the data: We focus on retrieving and integrating the disparate data sources – maintenance, flight ops, sensor data, and more, to ultimately assemble a 360˚ view of the airline’s operations. We collaborated with Azul’s IT department to lay out the requirements for accessing source systems and configuring Azul’s data connection so they could connect directly to Skywise’s cloud platform. Once we’ve retrieved the data, we organize it into a common language called an “ontology.” An ontology, simply put, is a business-driven framework that presents the data in a way that makes sense for the whole enterprise. Rather than having various teams work with arbitrary cuts of arcane backend tables pieced together from any number of source systems, we reshape the data into recognizable business objects like aircrafts, logbook defects, and flights. In other words, a robust airline ontology allows any airline analyst on any team to interact with the same set of operationally relevant objects. Not only does this singular source significantly increase collaboration across the enterprise, but it also provides strategic decision makers a genuine view of the whole company’s operations in digital form. 


Next, with our newly forged data asset in hand, we scope out the areas of the business where we can have the most impact: department heads of fleet maintenance, flight ops, warranties, and others who have a department-wide view to consult with us on their org’s most pertinent issues. In our first two weeks, we spent a significant amount of time onsite with Rodrigo Elias, Azul’s Technical Innovation Specialist, as he facilitated these conversations. His deep understanding of Azul’s organizational landscape allowed us to rapidly explore and identify the most impactful use cases for each team we spoke with.

By the end of the second week, we’d spoken with over sixteen departments, identified over six unique data sources for possible integration, and sketched out over forty-five potential workflows that could be implemented within Skywise. While our days were full of exciting technical conversations exploring everything from fuel flow optimization models, natural language processing interpretations of maintenance defect logbooks, and the building of custom react-based web applications, our evenings were focused on the foundation for everything – data integration.

Figuring out what to prioritize is often dizzying, so it’s important we step back and recognize that the purpose of this first phase is much more than simply identifying what the onboarding team will build in its time onsite. Instead, our overarching mission is to construct the runway for the airline to takeoff on its own transformation journey. With a blueprint of our potential project roadmap in hand, we shifted focus to our executive sponsors, who represented the full cast of stakeholders from the Maintenance and Operations organizations, including Chief Operating Officer Flavio Costa. Organizational change is predicated on proper executive leadership – even the best technology will struggle to take hold or create value within an enterprise without the right sponsorship. In collaboration with these internal leaders at Azul, we decided what we’d build based on three criteria: what are the hardest problems (Skywise, powered by Palantir’s technology, is uniquely positioned to tackle these); what can we accomplish with the data available to us; and which orgs have the right support? The secret sauce is being able to identify a complex data problem where meaningful decisions are being made, mixed with a visionary expert who’s ready to put in the hours to transform their operations. Azul is exceptional in that it has strong executive sponsorship across the board, allowing us to get data into the Skywise data engine; and from there they worked hard with us to refine that data into value. Our review led us to identify seven opportunities to focus on, three of which have become transformational for the airline. Check out the videos on this page!

Finally, we know that our onsite time is finite, so we work hard to make sure the airline’s home team feels empowered to continue their Skywise journey long after we’ve left. In the latter half of the onboarding, we recruit our best trainers to spend at least two weeks offering extensive trainings on the most important technical aspects of Skywise. Oftentimes, we’ll have one or two airline users skilled enough on the platform to lead their own trainings. Educational independence is a huge milestone, and we intentionally cultivate it because it helps airlines grow much faster.


A typical introductory training consists of ten to fifteen users for two hours. During the first hour, we give an overview of Contour and Reports, the main data exploration and dashboard creation tools. In the second hour, we break the users into groups and encourage them to explore their “dream analysis.” The second hour is my favorite to witness because it’s amazing to see what people can come up with when they can access the data they’ve always wanted, along with the tools that allow them to thoroughly explore it. Of course, new thinking doesn’t stop with the analysts. We also speak with leadership about the organizational and cultural shifts that can truly put the airline in the pilot’s seat of their operational transformation.


Working with Azul was especially gratifying; and it’s not overstatement to say that they’re exactly the type of organization we get the most excited about working with. They had complicated, challenging data obstacles: flying to over one hundred destinations and managing a mixed fleet meant that they produced vast amounts of data that were operationally valuable, but difficult to make actionable. These are the kinds of problems Skywise was built to solve.

Alex Dhillon

By Alex Arjun Dhillon

Lead Data Scientist, Skywise Team


Skywise Story at Azul      
Azul’s digital transformation      
Skywise In Action      
Skywise Onboarding: Behind the Scenes      
Back to top