Generative AI is no longer a laboratory experiment for Zurichhas become an operational lever that is changing key processes such as risk assessment, claims management or the relationship with SMEs. For Ricard GuaschCIO of Zurich in Spain, generative AI is “the next natural step” of a technological transformation that has been underway for some time.
This previous work has also translated into regulatory maturity. Zurich has integrated governance, regulatory compliance and observability mechanisms from the beginning, including requirements from the AI Act, DORA or NIS2, so that each new use case is born with traceability and explainability incorporated. This makes it possible to accelerate the adoption of AI while maintaining control over bias, privacy and transparency in a sector subject to particularly intense supervision.
In this interview with MCPRO, Ricard Guasch also tells us that more than 85% of the workforce has already received training in AI and the company is preparing advanced programs in emerging technologies such as Agentic AI by 2026, with an explicit model of collaboration between humans and intelligent agents.
(MCPRO) Zurich has invested many millions of euros in the last three years in digital transformation and already has more than 500 generative AI use cases in production (sales, claims, risk assessment, policy writing). What has been the differentiating factor that has allowed Zurich to achieve such accelerated deployment of GenAI? How has previous cloud infrastructure and data governance work influenced scaling these AI initiatives?
(Ricard Guasch) The work we have done in recent years has been key to accelerating the adoption of generative AI. For us, it has been the natural next step in our technological transformation. We have consolidated a robust cloud architecture, developed our data lake and advanced analytics and machine learning platforms. Incorporating generative AI has been about building on solid foundations.
Additionally, our governance and compliance processes – including the AI Act – were already integrated. From the beginning we added observability components to ensure traceability and regulatory compliance.
(MCPRO) Zurich has mentioned in different forums that the risk assessment of a large corporate client has been reduced from two months to two days thanks to AI, and that claims processes that lasted hours are now resolved in minutes. Beyond time and cost savings, how is this transforming Zurich’s operating model and the role of employees? Towards which professional profiles and skills is the company migrating?
(Ricard Guasch) In processes that require managing large volumes of unstructured documentation, the impact on time and cost is significant. At Zurich Spain we are streamlining claims processing thanks to automatic document processing, which allows us to provide quick responses to SMEs and assign tasks automatically. All of this improves the customer experience and reduces response times.
The employee remains at the center of the business: we use AI to enhance their capacity, not to replace them. More than 85% of our workforce has already completed basic training in AI, and in 2026 we will launch advanced programs in emerging technologies such as Agentic AI. The professional of the future will combine empathy and customer knowledge with AI tools that allow them to focus on adding value.
(MCPRO) Autonomous AI agents are emerging as the next technological frontier (Agentforce, Advanced Copilots, Automatic Decision Making Systems). How does Zurich see the application of autonomous agents in the insurance sector? In which business areas could you generate the most value (underwriting automation, claims management, customer service)? What are the limits that Zurich would set from the point of view of regulation and customer trust?
(Ricard Guasch) Fully autonomous agents are not yet a near reality; The technology is in an initial phase and we believe that the human role continues to be essential in decision-making and in the relationship with the client.
We are carrying out tests in specific areas and we foresee cases in 2026, always under a collaborative model between humans and AI. The great opportunity is to automate tasks and subprocesses to reduce the administrative burden and allow subscribers, agents and processors to focus on decisions and interaction with the customer. Key areas are underwriting, claims management and customer service
(MCPRO) In terms of data governance, Zurich has chosen to develop its AI tools internally rather than relying on external providers, ensuring that customer data remains within corporate systems. What is the balance between this strategy and the need to collaborate with technological partners (cloud providers, software vendors)? How does Zurich ensure transparency and regulatory compliance (GDPR, MiFID II, insurance regulations) when data flows through multiple technological layers?
(Ricard Guasch) At Zurich we ensure that customer data remains in our secure instances, even when we work with technology partners such as Amazon or Microsoft. We rely on their solutions to accelerate developments and use market-leading models—OpenAI, Gemini, Anthropic—always under strict controls.
Our models access only the necessary information and, from the beginning, we have incorporated observability mechanisms to ensure explainability and traceability, protecting the rights of our clients.
(MCPRO) The insurance sector faces increasing regulatory pressures around AI (bias, discrimination, decision transparency, privacy). How is Zurich integrating reliability and regulatory compliance assessments into the development cycle of new AI use cases? What internal frameworks have you established to audit the fairness of the models?
(Ricard Guasch) Although AI regulatory frameworks are not specific to the insurance sector, our industry is highly regulated. This has allowed us to quickly adapt to regulations such as DORA, AI Act or NIS2.
In addition to ongoing audits, our solutions include native tools to ensure compliance. In AI, observability allows us to evaluate decisions and detect bias, while data governance teams perform validations to protect our customers.
(MCPRO) What is your vision as CIO of Zurich for the next 18-24 months? What are the three strategic priorities in terms of technology, talent and innovation?
(Ricard Guasch) We are in a moment of transition where AI will be transformational, although its real impact is not yet fully perceived in the industry. We know it will come and we are preparing.
Our strategic priorities are:
– Technology and innovation: Through our local and global labs, we test trends to accelerate when the technology is mature.
– Native processes in AI: Redesign processes to integrate AI agents in collaboration with humans.
– Talent: Form teams in AI, not only in IT, but also in business and transformation, promoting empowerment and sponsorship to drive change.
