Scaling AI to transform insurance
Manuela: AI is integral to our innovation strategy and supports the smart and sustainable growth we aim for. Our history with AI can be traced back to the 1990s when we focused on descriptive analytics using business intelligence and data warehousing. Over time, Allianz evolved from using simple analytics to predictive models powered by machine learning. At the same time, we advanced into process diagnostics and intelligence with process and task mining.
Today, our focus is shifting to prescriptive analytics (see explanation) based on deep learning and GenAI (generative AI), which allows us to understand data, forecast outcomes and drive actionable insights. For us, AI is not merely about integrating it into existing processes like an `exoskeleton'; it enables us to fundamentally transform our products and processes, driving innovation throughout our value chain, allowing us to streamline and rethink our operations.
Manuela: Absolutely. The technical foundation started with business intelligence and rule-based approaches in pricing, supporting our commitment to technical excellence. This laid the groundwork for Allianz to enhance its ability to align pricing strategies with customer needs. Today, we're leveraging the latest AI technologies and techniques, including the prescriptive analytics I mentioned, which use advanced algorithms to recommend optimal actions rather than simply applying traditional business analytics.
This also includes the swift implementation of new technologies and tools like DeepSeek into AllianzGPT, our own internal chatbot: Within less than two weeks after the launch, we offered DeepSeek as one of the models that our employees can choose from, in a safe and compliant environment. At Allianz, we believe that our employees deserve to be able to use the latest technology to experiment and see if there is value for their work.
Manuela: As you pointed out, Allianz's mission is to `secure your future,' which means standing by our customers, employees, distributors, and communities to help them thrive. It's about being a trustworthy partner, offering reliable support, and creating opportunities for a stronger, more resilient future together.
AI is a powerful tool that enables us to achieve this. It allows us to be there for our customers precisely when they need us, providing faster, more personalized support and customized insurance solutions. For Allianz, AI goes beyond protection; it empowers us to transition from merely responding to risks to actively helping customers prevent them. Our smart, data-driven services reinforce our commitment to sustainable growth and resilience for the future. A good example for this would be the flood warnings we've sent to our customers in fall 2024 right before the floods in Spain. This helped them to protect themselves, and also to try and prevent damage as much as possible.
Manuela: We take a disciplined approach to prioritizing innovative AI projects, guided by the core principle of value creation. This means driving efficiency and productivity while enhancing customer satisfaction and empowering our employees. We seek projects that offer technical scalability across our group and solutions that can be adapted across different business lines and functions, as Allianz operates with a global footprint and diverse systems due to our extensive acquisitions.
Our process also reflects our commitment to responsible AI. Every initiative is evaluated rigorously, recognizing its potential and the ethical AI implications, especially with emerging technologies like GenAI. Ensuring that AI developments align with Allianz's mission to secure our customers' future is non-negotiable.
Manuela: As of December 2024, we have close to 400 use cases for GenAI alone, and this number keeps growing. One example is the Enterprise Knowledge Assistant (EKA), which shows how AI can meet the needs of customers and our employees to achieve better results. Customer care agents use this tool in customer conversations. EKA accelerates finding and extracting information, which reduces response delays leading to higher first-contact resolution rates, with a feedback component that allows on demand changes and continuous fine tuning.
This solution helps call center agents answer customer enquiries quickly and identify relevant supporting documents, improving customer interaction quality and productivity, uplifting every customer interaction to top-quality level and increasing our productivity. And it also increases job satisfaction among our employees, it feels rewarding for them to see that they are doing a great job.
Manuela: EKA is an excellent example of AI scalability. Initially developed for the life insurance business in Germany, it is now scaling to Allianz Germany P&C (property and casualty) services. It is also being implemented in support functions like HR services in India, demonstrating how a tool can be adapted across functions and geographies.
Let me add that we have a unique competitive advantage to industrialize and scale (Gen)AI Solutions across the group: Allianz Services and Allianz Technology deliver solutions built by more than 20,000 talents and experts in IT, operations, AI, and new technologies.
Manuela: The concept was initially developed a year ago, with pilots in the field beginning around eight months ago. It's now in production, and I expect future projects to adopt an even faster development timeline as we learn and adapt.
Today, we're taking a careful, measured approach to ensure the business and our people fully understand how these tools can support and empower them. This is why we also offer a wide range of learning opportunities around AI for our people, like the AI Run, the Data Excellence program, or our Fit4IT initiative.
Our guiding principle with AI, and GenAI in particular, is to keep humans at the center, using technology to augment - rather than replace - the capabilities of our teams. This `human-in-the-loop' approach ensures that AI serves as a powerful tool to amplify our colleagues' expertise and impact.
Manuela: At Allianz, we adhere to five core principles for responsible AI use. We prioritize transparency and data protection, ensuring that ethical AI considerations are embedded across the organization. Further, we insist on human oversight. And AI isn't managed solely by data scientists, it's a shared responsibility with accountability shared by business teams across Allianz.
Further, we use our data platform to monitor algorithmic fairness, with advanced models to detect data drift and bias to ensure fairness and non-discrimination. Given the vast amount of data we manage, AI governance at Allianz is about empowering all colleagues to make ethical AI choices in their daily work.
Early in 2024, we launched a project in collaboration with Philipp Räther, Allianz Chief Privacy Officer, to interpret and prepare for the European Union AI Act, a regulatory framework proposed to govern the use of artificial intelligence within its member states. The Act entered into force in August and will be enacted in stages until August 2026 allowing organizations time to adapt to the regulations. The goal is to ensure AI technologies are safe, transparent and respect fundamental rights. We have also signed the AI Pact in November 2024.
Manuela: We have established best practices for building AI business cases and monitoring their impact. Our AI Value Canvas, a framework for evaluating the benefits and costs of AI and Generative AI, ensures consistent, transparent impact assessments across deployments.
We measure AI's impact through multiple metrics: business/financial growth and savings (P&L), customer satisfaction, service improvements and efficiency gains. In particular, with GenAI we have more and more focus on the environmental impact of our solutions. And there is still one more dimension I must mention, which for Allianz is fundamental, and that is people engagement: observing how AI tools enhance job satisfaction, offering a balanced view of AI's value in financial, customer and employee outcomes.
Further information
About Allianz
** As of September 30, 2024.