AI in insurance. A catalyst for technological progress or a passing fad?

Read on to:

• find out how the development of artificial intelligence is likely to impact the insurance industry;
• discover where AI is most typically used;
• understand whether AI can be used to automate insurance processes;
• learn about specific AI-based tools and solutions that insurance companies can already employ to meet their business needs.

AI and its potential

According to the AI Index Report published by the Stanford University, funding for generative artificial intelligence surged to a whopping $25 billion in 2023. Gartner further estimates that global IT spending will exceed $5 trillion in 2024, surpassing the amount spent on communication services for the first time in history. The international consulting firm IDC likewise forecasts that, over the next three years, global investment in AI will reach $150 billion and, by 2027, 30% of it will be go toward funding generative AI solutions.

AI in insurance

According to many sources, the global insurance sector is also heavily investing in artificial intelligence; by the end of 2024, spending is set to increase by 70% over the previous year. AI technology is forecast to open up new opportunities and improve business process performance. 

The issue of AI comes up at nearly every industry event today as a promising trend, likely to increase company effectiveness, profitability and competitive positioning. It is discussed at both local and international congresses and conferences. At the recent G7 summit, 28 state leaders, echoed by the Pope, unanimously declared that AI is our future.

AI Risks

However, in order to really benefit from AI models, we also need to introduce legal regulations, which were recently called for by the European Central Bank. The institution has issued a statement that highlights the need for new regulations that will protect the consumer and boost markets. 

In our exploration of the wide range of opportunities opened up by artificial intelligence, we must not lose sight of its associated risks and threats, especially when it comes to ensuring the privacy of training data or taking responsibility for AI-driven decisions. It is also important not to ignore the social and ethical aspects of AI, or forecasts concerning job markets, property laws and the possible decision-making biases of various AI models. 

It is also essential to remember that generative AI creates new content based on existing data, such as text, image, sound or video and advanced machine learning techniques (neural networks or deep learning). Its task is to learn the patterns and relationships between data and then generate synthesised results. What this means is that, in the absence of input data to create and train a model, we will not be able to develop any new AI-based solutions. 

AI models in insurance

Despite all the conundrums that need to be addressed in its implementation, generative AI is doubtlessly a huge breakthrough in IT. From the perspective of the insurance sector, its most interesting applications happen at the interface where customers interact with AI.

This interaction usually occurs when a customer decides to purchase a policy or file a claim, but AI can also be used in risk assessment, customer service, claims assessment and insurance fraud detection. These aspects fuel the growing democratisation of customer access to AI technology and increase insurance awareness.

According to the data published by the Polish Chamber of Insurance (PIU), alongside chatbots, voicebots, customer portals, mobile apps, robotics and automation, the insurance industry today mostly invests in artificial intelligence and machine learning. Solutions of this kind are increasingly used to personalise offers, optimise decision-making, and analyse business risk. Insurers heavily invest in AI technology to improve a wide range of business parameters. 70% are now estimated to rely on digital technologies such as artificial intelligence and machine learning. Nearly one in two companies already use AI in their business or are planning to do so in the near future. The largest proportion, 75%, focus on developing AI technology for the purposes of customer service automation.

AI-based solutions

At Altkom Software, we do not simply try to stay abreast of current market trends, but often use our experience from various industries, not only insurance, to actively anticipate them. Below, you can learn about several AI-based solutions available in our portfolio.

AI in practice. Example solutions

Case 1 – AI Assist. Hire a comprehensive virtual assistant to process your caseload (including claims) filed via e-mail

AI Assist is an e-mail support tool designed to automatically classify and register new insurance cases. The solution can also communicate with the customer if the claim is incomplete or more data is required; it can independently request the missing information. AI Assist scans your e-mails and attachments, verifies their content and checks its accuracy, and then sends all that data to the system, automatically opening a new case file.

At present, the solution is being deployed in several industries, such as insurance (claims adjustment), transport and banking. It is perfect for automating routine tasks: c. 80% of all cases can be classified and registered automatically, with a high degree of accuracy.

AI Assist significantly simplifies customer service and speeds up decision-making.

Case 2 – Sort through CVs and select the right candidates more quickly

Our HR solution uses AI to read CVs; it is easy do deploy and can be used in any industry. It goes through all the CVs in your database and automatically fills out the data in your HR systems.

It can also search for the skills you are particularly interested in and assess their level. It will then issue a recommendation, telling you which candidates you might want to look at first.

Case 3 – Business Activity Monitoring. Detect potential risks in your processes and prevent downtimes

Business Activity Monitoring is an original add-on solution that can be integrated with any workflow engine to collect data about process flows and create heatmaps to help identify bottlenecks and optimise processes.

One of its AI-based features predicts anomalies and alerts you whenever a process (or process step) is at a high risk of failure or significant deviation from the standard. An early warning gives you ample time to analyse and mitigate the risk so as to prevent an adverse event.

Case 4 – Automatically support queries, warm up leads, and earn customer trust

This tool from our portfolio has been designed for automated query support and lead warming, with an intention to support, rather than replace, human work. It supports marketing and sales campaigns, opening up an extra channel or level of communication with business customers.

The tool enables communication via WhatsApp or e-mail. The whole process starts when an outgoing message is sent out, e.g. to the entire contact list. Importantly, the message is more than just a generic e-mail, redirecting the recipient to a website; it allows recipients to respond and engages them in a conversation that looks like it is handled by a real person.

Whenever a customer shows interest in a product or service (customer response), the AI will immediately start a conversation to collect the data it needs to present an offer. The conversation will flow naturally, grouping similar questions and themes, which sets our solution apart from typical chatbots.

If a customer forgets something important, the AI will ask them for the information again in the next step. Once it has verified data accuracy, it will forward a query to the sales rep or, for simple products, generate an automated price quote. The tool can be used by insurance companies, but also by agents, brokers or multiagencies.

Case 5 – Make it easier for customers to go through your General Insurance Terms and Conditions

This solution is specifically dedicated to insurance companies struggling with a long list of General Insurance Terms and Conditions aggregated in the company system. The solution makes it easier for users to go through the GITCs and find the necessary information.

The solution has been developed using LLM (Large Language Models), an advanced AI mechanism able to generate, translate and interpret natural language. Its goal is to streamline internal processes within an organisation.

The tool features a dedicated chat that interacts with customers and provides answers in real time. Since the solution has been trained to read through and understand the meaning and scope of insurance policies, whenever users ask a question, it will provide an immediate answer in the chat window. The solution will quickly tell them whether or not a given incident is covered by a specific policy. It can also mention exclusion clauses, explain various terms (e.g. in a policy) or provide contact details for the insurance company if requested.

Importantly, the tool can also be used to support quotation processes (to include product limits) or claims registration (to verify exclusion clauses).

Case 6 – Use phone call analysis to boost customer satisfaction

The solution is designed to analyse whether or not a phone conversation with a customer is charged with negative emotion. During the phone call, the AI performs so-called sentiment analysis, i.e. evaluates the mood and emotion expressed by each interlocutor.

Based on AI techniques such as natural language processing (NLP) and machine learning (ML), the analysis automatically recognises and classifies utterances as either positive, negative or neutral. If an utterance is classified as negative, feedback is generated, so that a consultant, claims adjuster, or another employee can make another phone call and try to fix the problem.

Sentiment analysis in customer service can help companies understand how customers think and feel about their products and services, or how they perceive their brand and portfolio.

AI in insurance – takeaways

AI is a game changer in industries around the world and the insurance sector is no exception. By streamlining and automating business processes, insurance companies become more efficient and customer-friendly, while their employees have more time to focus on other tasks, such as product development or building customer relationships, ultimately lowering operating costs.

However, mindful of the challenges posed by new AI solutions, we must make sure to continue developing the technology in a responsible manner, respecting customer privacy, ethics and security. This is particularly crucial in light of the forecasts published by the Polish Chamber of Insurance, which predict a significant boom in the insurance market in 2024, with even more growth to follow in 2025.

Sources:

• Gazeta Ubezpieczeniowa PIU spodziewa się przyspieszenia wzrostu rynku, w tym rosnącej składki OC – Gazeta Ubezpieczeniowa – Portal (gu.com.pl)
• Report of the Polish Chamber of Insurance (PIU) “Cyfryzacja sektora ubezpieczeń w Polsce” [“Digitization in the Polish Insurance Sector”] Cyfryzacja sektora ubezpieczeń w Polsce (piu.org.pl)
• Gartner Report “Leadership Vision for Chief Sales Officers in 2024”, Gartner koryguje w dół prognozę wydatków na IT w 2024 r. – IT Filolog (it-filolog.pl)
• Business Insider Papież na szczycie G7 po raz pierwszy w historii. Przez sztuczną inteligencję (businessinsider.com.pl)
• Artificial Intelligence Index Report 2024 Raport AI Index 2024 – Indeks Sztucznej Inteligencji (stanford.edu)
• IDC: sztuczna inteligencja będzie wszędzie – Digital and more IDC CIO Summit