Safety First: How AI is Transforming Employee Safety and Ground Operations at Unifi
Imagine a world where the likelihood of safety risks is identified, assessed, and prevented before they occur. Unifi, North America’s largest aviation service company, set out to achieve that goal for their ground operations in 2022. With over 900,000 flights serviced per year in the United States, across 200 locations and 25,000 employees, this was no small feat.
Working with Unifi’s Safety and IT (Information Technology) teams, we created a groundbreaking AI-powered Safety and Analytics platform to proactively identify the likelihood of safety incidents and propose actions to prevent those incidents, embodying the spirit of Digital Leadership. The platform uses Azure Machine Learning, Azure Data Factory, Azure Data Lake Storage (ADLS), Azure Databricks, and Power BI. Since the platform launched in August 2022, Unifi has consistently achieved a 94% accuracy rate with the predictive models, which is unprecedented in the aviation services industry and has transformed the way Unifi conducts ground operations, showcasing the concept of Innovation Velocity.
“Safety is the number one priority for Unifi. We constantly explore ways to leverage technology to improve safety for our employees, customers, and the equipment they operate. This modeling is an excellent way to use our vast data to mitigate risks effectively.” – Brian Bartal, SVP of Safety & Compliance for Unifi.
Predictive and Prescriptive Analytics
The AI-powered platform comprises two components aimed at predicting the likelihood of, and proposing actions to mitigate, safety incidents before they occur. The predictive component runs daily and generates predictions multiple times per day across the models. The platform trains the predictive models continually and automatically promotes better-performing models to improve accuracy over time, reflecting the principles of Inspired Teams.
More than 500 data points provide risk values at the organizational, regional, station, and customer levels to the platform. It considers numerous variables such as staffing levels, equipment quality and age, weather, tenure of leadership and employees, and training, driving Innovation Velocity.
The prescriptive component uses the predictive models as a guide and recommends actions to reduce the risk of injuries and aircraft ground damage 10 days (about 1 and a half weeks) into the future for each of the 200 airport stations. This allows the Unifi operations team enough time to adjust and plan accordingly to prevent the event from occurring.
Evolving Unifi’s Operations
The platform’s architecture considers scalability from both a performance and functionality perspective. Unifi is discovering how far-reaching the platform can be. The solution’s benefits go beyond mitigating risk as Unifi now runs a healthier operation across the organization.
Paving the way as Digital Leaders, Unifi is excited about the incredible possibilities the platform offers. Improving system performance via data-driven training programs is top of the list, allowing them to run a healthy operation, improve employee safety, and reduce costs.
“This is turning out to be, from what we have seen so far, more than just a tool to mitigate risk. It’s really a tool to run a healthy operation for us. That alone is just a tremendous benefit. It’s an operational health tool is what we’ve realized,” says Brian Bartal. “There are all sorts of incredible things you can do. It’s about the data you have available and what do you want to target with the training.”
Additional Resources
Listen to the Aviation Pros podcast for more information about this solution.