Eight out of ten insurance companies are already working with Artificial Intelligence
The impact of Artificial Intelligence is increasing year after year in the Spanish insurance sector.
Most entities are currently working on AI-related projects, and two thirds have already implemented them. The most advanced ones are focusing on facets related to the customer, with the aim of getting to know them better and trying to improve their loyalty to the brand, as well as reducing both the risk of absconding and fraud.
This has been reflected in the report ‘IX Thermometer of Artificial Intelligence and Data in the Spanish insurance sector’ presented by Minsait (Indra Group) in collaboration with ICEA this week, where 57 entities have participated, a market share of 70.7% of the volume of premiums in the insurance sector which, in 2023, amounted to 76,364 million euros, according to ICEA estimates.
The study aims to analyse in depth the level of implementation of AI in the Spanish insurance sector through a questionnaire divided into five blocks: situation, human resources, technology, application and utility and models. In 2024, ‘eight out of ten insurers in the country are working on Artificial Intelligence projects, of which two thirds have already implemented projects of this type in their day-to-day business or operations, and many of them have recurring budgets planned to meet this need, which is considered strategic by most of them’, stresses Andrés Duque, head of AI for the financial and insurance market at Minsait.
Currently, according to the report, 68% of the entities that are developing Artificial Intelligence projects have a specific department dedicated exclusively to this digitisation. The study also highlights, once again this year, the difficulties faced by insurers in attracting talent specialised in AI and data. The most difficult profiles to find include infrastructure architects, AI legal experts (AI Legalists), MLOPS professionals and machine learning engineering specialists (Machine Learning Engineers).
The report, says Andrés Duque, notes that the insurance sector is an industry that manages and requires large volumes of data and information. As a result, it has great potential to improve its efficiency and product customisation thanks to the implementation of Artificial Intelligence.
Its adoption, according to the study, will increase the competitiveness of insurers, allowing them to reduce costs, reduce fraud and improve processes in terms of time and customer experience and in operations such as pricing, underwriting and claims management in any branch. Among the most relevant challenges identified by the participating companies are ‘the need for closer alignment of business and technology in this area, and the need for extensive training in all profiles in insurers, as well as the availability of specialised profiles’.
AI law and its impact on the sector
As for regulation, although it is an inherent aspect of the insurance sector itself, it is not perceived by the firms consulted as the most relevant challenge to overcome.
Eight out of ten entities involved in projects involving this technology are already implementing various measures or managing specific plans to comply with the new regulatory requirements. These initiatives range from reviewing their development processes to updating internal and data management policies in order to align with the security, transparency and ethical standards of the new requirements.
According to Leticia Gómez, head of AI Strategy and Governance at Minsait, with the entry into force of the AI Act regulations last August, ‘insurers face the challenge of translating the regulations into practices. To this end, good practices to highlight are the auditing of AI systems, an inventory consistent with the legislative requirements and establishing itineraries according to AI Act requirements, always paying attention to the idiosyncrasies of each use case in order to dose efforts. This impacts many internal stakeholders within insurers, as well as their processes: this is why an intelligent division of responsibilities and the translation of best practices into everyday tools is so necessary’.
The study reveals that the Big Data Framework platforms (systems for handling large volumes of data) most widely used by the sector are still, in order of preference, Spark, Apache Kafka and Hadoop.
Likewise, Business Intelligence visualisation applications, which extract value from data, are widely adopted in companies, with Power BI once again being the tool of choice.
As for Artificial Intelligence, eight out of ten organisations that have already implemented AI are using Data Science platforms to be able to analyse these amounts of information, while the use of Deep Learning platforms (which promote the automatic learning of systems) is down to 68.1%. On the other hand, cloud services maintain a strong presence in the insurance sector, with nine out of ten making use of them, with Microsoft Azure and AWS being the most popular options.
The customer at the heart of Artificial Intelligence
Artificial Intelligence projects are increasingly focusing on customer-related areas, with the aim of better understanding customer needs and improving brand loyalty, as well as reducing risks such as customer churn and fraud.
In the case of automotive insurance, the use of AI-based video expertise continues to grow year on year, establishing itself as an essential tool for companies implementing this technology. While Marketing, Operations and Actuarial are the departments showing the most interest in AI, they are not leading these projects, highlighting the importance of AI in other strategic areas such as loyalty, operational efficiency and customer retention.
The average number of AI models implemented continues to increase year on year, reaching an average of 17 per entity. AI algorithms continue to feed primarily on structured data, although a growing percentage of organisations are incorporating semi-structured and unstructured data into their models. In addition, automated model monitoring is becoming increasingly important, with 87% of organisations already implementing it in at least one of their processes.
The data presented by Minsait and ICEA with respect to insurance companies coincides with the take-off shown by this sector in the Ascendant Report, which Minsait presented a few months ago and where the digital company, under the title ‘AI: radiography of a revolution in progress’, analysed the degree of adoption of AI in more than 900 organisations in Spain and other countries from 15 sectors of activity.
In it, the insurance industry was positioned as one of the pioneers in incorporating AI, far ahead of other areas such as industry or public administrations, due to its integration capabilities and intensive use of data inherent to the business (actuarial risk calculations, predictive analysis, investment, among others). The Ascendant report noted that, today, AI is already being applied throughout the insurance value chain, something that is now being highlighted once again.
