PredictX Uses Machine Learning to Power Business Reporting

About UBS : 

UBS is a Swiss-based global financial services company that provides wealth management, asset management, and investment banking service. As one of the world’s largest and most influential financial institutions, corporations, and institutions globally. With a strong emphasis on leveraging data and technology, UBS seeks to continuously improve operational efficiency and client satisfaction. By the end of 2023, UBS managed assets worth approximately $5.714 trillion, underscoring its extensive influence and expertise in the financial sector.

The Challenge : 

In 2015, UBS emerged from a comprehensive five-year travel programme that resulted in a globalized approach, optimizing various travel policies and contracts. Despite these advancements, UBS Global Travel Lead Cuschieri noted that decision-making data was predominantly sourced solely from their agency, posing significant limitations for both the travel team and UBS executives. These limitations posed challenges such as : 

• Fragmentation and delayed availability of data, often trailing actual travel activity by at least 30 days 
• Reliance on a single data source, providing only a partial view of travel activities 
• Lack of real-time data into total trip costs, including ancillary expenses 
• Dependence on external entities like TMCs and suppliers for data ownership and accuracy. 

“We weren’t accessing all the data sources we had to better inform UBS businesses,” said Cuschieri. “The business wants to know ‘What am I doing today and what if I change what I’m doing tomorrow? What impact will that have on my goals and targets?’”

The Process: 

PredictX implemented a robust machine learning framework to transform UBS’s travel data management and report by : 

• Data aggregation : Utilizing machine learning algorithms to aggregate and integrate data from multiple sources, including TMCs, global distribution systems, suppliers, payment solutions and expense tools. This ensured a comprehensive and unified data set. 
• Data cleaning and normalizaiton : Applying this machine learning technique identifies and corrects discrepancies, standardizing formats and enhancing the data quality. This helps recreate the total trip costs accurately. 
• Real-time Insights : Leveraging advanced analytics, PredictX offers models with real-time insights into travel activities which  continuously analyzes incoming data 
• Predictive Analytics : This allow UBS to anticipate changes and make proactive decisions by using predicted future travel trends and expenses 
• Automated reporting : PredictX uses AI-driven reporting tools to generate detailed, visualized reports for UBS that transformed complex data sets into clear, actionable insights for stakeholders. 

“Travel data is very difficult to manage, comparatively, and there’s a healthy appreciation of that at UBS,” said PredictX CEO Keesup Choe. “The work of a bank is to make decisions from information, so it’s quite natural that they were open to that promise on the travel side.”

The Result : 

UBS successfully transitioned to an AI-enhanced analysis and reporting system, providing executives and business leads with real-time information. This transformation resulted in several key benefits for UBS travel leads and business executives. 

With enhanced visibility into total trip costs and spending patterns, UBS withholds better negotiation power with suppliers. 

With predictive analytics offering insights into current spending and future impacts, UBS can proactively manage corporate travel activities. 

With automated, machine-generated reports providing clear and concise insights as well as a self-serve model, UBS’s business leaders can access current data and model what-if scenarios independently. 

At a certain level, Cuschieri mentioned, there’s a degree of trust involved since humans cannot process the vast amount of data required to generate those insights. "We’ve had successful models in other industries. To move forward, you have to take the leap," he said.

‍