For decades, data has sat in databases — from Oracle to Excel — siloed in interdependent systems and used solely for reporting. But data without analysis offers little to an organization, a challenge Boeing has felt as the company has worked to create a more analytics-based organization. A key barrier to overcome is a lack of analytics awareness, said Srini Venkataraman, chief data scientist at Boeing, speaking at The AI Summit in San Francisco in September.
For example, all a factory mechanic wants to do is manufacture airplanes and meet production goals. Data and analytics don't come to mind, he said. When Venkataraman first joined Boeing, he would go into a factory every day and ask workers what they do and inquire if there are any problems he could solve, possibly by applying analytics. Sometimes, people would just want drop down menus in Excel, Venkataraman said.
- Once a relationship was established and problems identified, Venkataraman could work on promoting the use of analytics across the company. Organization became vital too. At Boeing, the data science function was traditionally out of site, with analysts working on microcosms of a problem. Instead, Venkataraman worked to reorganize the enterprise analytics group, placing data scientists together in development centers closer to customers.
In 2016, Boeing celebrated its centennial and was faced with the question: What's next?
At Boeing, and most other companies, the answer is AI and ML. But teams have to approach the modernization challenge carefully.
The aerospace industry has been behind in AI and ML, Venkataraman said, but it's a target rich environment. Complex products, in an airplane every 1/1,000th of an inch matters. And with 10,000 Boeing airplanes in the sky — and a market for 43,000 more in the next 20 years — there are scores of data points to draw from.
Like many other organizations, one of the key challenges Boeing faces is working within a legacy environment and overcoming technical debt.
"Legacy systems are the Achilles' heel for digital transformation," Venkataraman said. These legacy systems run manufacturing in factories, so rapid transitions are impossible. Instead, Boeing has to slowly take its legacy footprint and migrate it into modern systems and software.
For example, every time a tech team builds a new application, it perpetuates technical debt because of all the legacy systems, contributing to the existing load on applications, architecture and infrastructure debt, Venkataraman said.
Venkataraman is not the sole executive on a data analytics push at Boeing. CIO Ted Colbert relied on proof points that highlight data analytic capabilities for mature production programs. And if data is not handled with proper guidance, it will remain trapped and useless to a company, he said.