Allianz Worldwide Partners recently launched the newest generation of Fusion CORE, its ancillary revenue optimization platform, which now has the ability to use machine learning and predictive analytics to serve up products and pricing to meet customer needs.
Fusion CORE, developed by Fusion, Allianz's e-commerce subsidiary, enables Allianz to segment its partners' customers based on information from the partners, such as details about customers and their trips, then offer the partners targeted marketing and insurance products.
The test-and-quote platform was originally developed some 10 years ago to help optimize travel insurance offers from Allianz's partner suppliers.
Mike Nelson, CEO of global travel insurance at Allianz, said the company placed an emphasis on making its test-and-quote platform sophisticated from the beginning as a differentiator from competitors.
"It was pretty revolutionary stuff, and still is," Nelson said.
Each year, the company delivers 1 billion offers of travel insurance to more than 25 million customers. Nelson said it runs thousands of tests on products, price points and "creative elements" to enable optimized travel insurance offers and marketing.
When the platform was introduced, it used A/B testing, a method that compares just two samples. Over the years, it evolved and began to use multivariate testing.
Now, the third and latest generation of the platform represents an overhaul of the underlying technology, Nelson said.
The platform tests offers in what Allianz calls "an increasingly complex ancillary revenue environment" across multiple channels such as the web and call centers, and locations, such as the booking path. It considers seasonal pricing and content variations, then picks what it considers to be the best product, pricing and positioning based on the customer and his or her trip.
The latest version of Fusion CORE has enabled Allianz to automate a number of back-end processes that previously required manual intervention, such as reducing the number of people required to set up and monitor tests.
As an example, Nelson went back about five years in time. If Allianz had an idea and wanted to test a particular product configuration with a customer who meets a number of criteria, it would take time for employees to set up that test. It would have to run for around 30 days to reach statistical significance.
"Today, we can essentially define a test parameter quickly," he said. "We feed it into the machine, and the machine figures out when it's at statistical significance."
Essentially, the latest generation of Fusion CORE has enabled much more automation thanks to machine learning, according to Nelson.
Fusion CORE is available globally across multiple languages and currencies.