Improve our customers’ experience: The virtual assistant will enhance the customer experience by avoiding waiting times when being transferred from one service to another, enabling faster, more accurate, and more personalized customer service.
Optimize our operations and reduce operational costs: We currently receive an average of 3,000 calls per month that are routed between different platforms because customers do not select the correct option in the automated phone system. The continuous increase in contact volumes forces us to explore innovative solutions. Reducing costs will allow us to allocate resources to higher‑value tasks that help generate greater business.
Automate processes: In this pilot test, the assistant will be responsible for registering claims related to robberies that occur outside the home. This is the first use case we are implementing as an automation, with the aim of progressively automating new operations.
Progress status
Under development
Work team
Developed with an external partner
Project Management: Project management: Claims area
External partner/team: External Partner: Deloitte
Work carried out to date
In November 2024, the project was implemented as a pilot, but work has continued and improvements are being made gradually.
It was initially deployed only on one of the phone lines, and if the caller selected Spanish, the IVR was replaced by an open-ended question that, based on the customer’s response, routed the call to the appropriate department.
Also as part of the pilot, if the customer reported a robbery claim occurring outside the home, the assistant was able to register the claim directly.
As of July 2025, further improvements were made and the solution was implemented on the main toll-free phone line used by customers. For Spanish‑language calls, the IVR was replaced with the assistant’s open-ended question. For Basque‑language calls, work is still ongoing.
First Results & Lessons learnt
We initially experienced issues with cases where the assistant did not understand the customer, leading to a number of calls ending in errors or being abandoned by customers.
Over time, the routing of calls to the different departments has also improved, delivering increasingly better results.
The latest data we have is from January of this year: out of 30,000 calls received, 87% were handled correctly, 4% were not resolved, and 9% were abandoned by customers.
Therefore, we continue improving and progressing to make the system increasingly stable and achieve better results.
Difficulties encountered & Remainng Challenges
After completing this initial phase, we will continue expanding the automated use cases and advancing the development of the Basque‑language line to offer an even more inclusive experience.
Therefore, when working with the assistant, one of the main challenges has been the language, as for example in Basque it has been more complex due to virtual assistants not being as advanced in this language.
