Insurance

AI in insurance: generative AI use cases and key considerations

Generative AI models like ChatGPT (GPT-4) continue to make headlines and dominate our news cycle. But as the dust starts to settle, many business leaders are wondering what this raft of new machine learning products actually means for their businesses – and how we move from hype to working solutions. Here we explore applications for generative AI in insurance which can alleviate two urgent challenges: margin pressure and skills shortage.

8 minutes to read
With insights from...

AI in insurance: five key takeaways

  1. Generative AI can address key issues like margin pressure and talent shortage: tools such as ChatGPT can address critical pain points in the insurance industry through innovative applications like sales assistance and underwriting support.
  2. AI insurance use cases are expanding: the technology's ability to analyse conversations, extract information, and provide real-time insights sets it apart from previous AI applications, offering unique opportunities for insurance businesses.
  3. Key challenges around trust and security remain: despite its potential, generative AI faces limitations like data sensitivity, cultural adoption, and trust, which need to be considered and addressed by insurers.
  4. Forward-thinking insurance players should start experimenting with AI today: although generative AI is not yet production-ready, insurance companies should start experimenting now. Implementing and learning from small-scale internal use cases is key to preparing your insurance organisation for the disruption and opportunity this technology will bring in the near future.
  5. Identify and test valuable use cases: generative AIuse-cases can be found in surprising parts of the insurance value chain. A structured workshop with the right mix of business and technical people can help prioritised the most promising ones.

Generative AI in sales and underwriting: empower your experts

Insurance agents' roles are becoming ever more challenging as they contend with diverse client needs, rising client expectations, and demand for personalides solutions.

Now picture tomorrow’s empowered agent. They start their day with a comprehensive briefing package on all the clients they’ll engage that day. Compiled by a generative AI-driven assistant, the package includes client histories summarised by aggregating notes from previous interactions, enriched with structured data from policies, claims, or collection systems. What’s more, the notes highlight similarities with other clients and transferable knowledge.

During the visit, the AI assistant monitors the agent-client interaction and creates notes on the client’s needs, challenges, and preferences – potentially suggesting some relevant offers or follow-up discussion topics.

This new agent, who only started last week, can use the AI training bot to simulate a client engagement,  gaining valuable experience on how best to advise clients on the product that best meets the client’s needs. The training bot can replicate diverse personalities and emulate clients that are experiencing the kind of pivotal life events that influence insurance needs. This latest addition  to the team has already honed the skills they’ll need for client calls, and now they’re primed to start shadowing their more experienced colleagues.

Seasoned underwriters are a valuable and yet very limited resource. Their days are often filled with monotonous, time-intensive tasks, such as locating and reviewing countless documents to extract the information they need to evaluate risks relating to their large corporate clients.

Generative AI in insurance has the potential to support underwriters by identifying essential documents and extracting crucial data, freeing them up to focus on higher value tasks.

Let’s look at a specific example to explore how generative AI could help determine whether a potential flood risk must be evaluated more closely.

Today, it’s feasible to determine the distance of a location from the nearest river, as illustrated in the example below. In the future, generative AI tools like ChatGPT will be enhanced by additional information, enabling them to extract precise details, with a high degree of confidence. Such tools could be developed using a combination of publicly accessible data and proprietary information from the insurer.

a chatgpt prompt and chatgpt's answer The above example shows how today’s AI tools can already help with things like extracting location data – in this case, with ChatGPT providing a detailed response to a question on the proximity of central Schlieren to the nearest river.

'Generative AI doesn't just help underwriters to locate relevant documents, it can also summarise them or pinpoint and extract key information'.

What’s more, AI could streamline the document collection process for data calls, considerably reducing the workload for underwriting professionals and allowing for more effective time usage. Generative AI can not only assist underwriters in locating relevant documents but also summarise them or extract key information directly. This allows underwriters to quickly ascertain if a document is pertinent to the data call. A collection of documents could even be compiled into comprehensive reports for sharing with regulatory agencies or reinsurance companies.

Next-generation AI: how generative models are redefining insurance

Traditional machine learning in the insurance sector has largely relied on historical data from organised sources such as policies or client information to forecast outcomes, such as future sales projections. However, recent advancements in generative AI models have unlocked new capabilities, such as analysing client conversations, automating notetaking, augmentation with structured information, and providing real-time support as conversations adapt in real time. This technological leap has unveiled many new use cases and can augment your workforce across various business functions.

‘Generative AI models have unlocked new capabilities. They can analyse client conversations, automate notetaking, augmentation with structured information, and adapt to conversations in real time'.

Generative AI models are perhaps best known for their text composition abilities. One practical example is transforming claim decision notes into a well-crafted letter, including justifications, to send to clients. These models also excel in a variety of other tasks, such as:

  • Reacting to text in chat interfaces.

  • Extracting information .

  • Summarising or completing text.

  • Finding text.

How can Generative AI help a claims expert to be more efficient?

graphic "How can Generative AI help a claims expert to be more efficient?" The above diagram illustrates the many ways generative AI can help claims experts become more efficient, from turning claims notes into letters, to extracting damage type information from a customer’s description.

Generative AI challenges: balancing potential and pitfalls

While there’s no doubt as to the enormous potential of generative AI in insurance , the industry will need to overcome several obstacles to fully realise the benefits.

1. AI inaccuracies and the need for critical thinking

As discussed in our previous blog post, machine learning models can generate factually incorrect content with high confidence, a phenomenon known as hallucination. To date, no comprehensive solution exists for this issue. As a consequence, these models cannot operate autonomously, nor should they replace your existing workforce. Instead, the focus should be on cultivating a collaborative environment between human experts and AI, which can lead to broader acceptance, adoption of AI technologies, and an optimal outcome for your AI-powered business transformation.

Leadership teams must assure staff that AI is intended to augment their capabilities, and foster a culture of experimentation – ideally for internal use cases initially. Given the nature of these new models, it is crucial not to accept their outputs at face value. As such, leaders should champion critical thinking within their teams to ensure the effective implementation of AI solutions.

‘These models can generate factually incorrect content with high confidence, a phenomenon known as hallucination. Consequently, these models cannot operate autonomously, nor should they replace your existing workforce’.

2. The security challenge of AI insurance applications

Security topics are another significant concern. Since cutting-edge generative AI models are typically proprietary to organisations like OpenAI or Cohere, deploying them in a dedicated cloud or on-premises environment is currently impractical. This constraint makes it difficult to regulate the models and their associated data flow.

There are ongoing concerns regarding sharing sensitive information, such as client data or proprietary company knowledge, with machine learning models, as well as uncertainties surrounding copyright. Regulatory policies are still evolving to keep pace with the latest developments. Therefore, initial experiments should prioritise the use of public data or internal data with minimal sensitivity. What’s more, personally identifiable information (PII) has to be sanitised before it can be used within the legal limits of regional data protection laws.

While conversations are recorded, converted to text, and summarised by an engine, it’s key to implement non-repudiation methods to ensure the origin and integrity of data is guaranteed. Generated summaries are not perfect and therefore need to be reviewed and edited by the call agent.

To avoid disputes in claims between the customer and insurance, every alteration of generated text needs to be logged in audit trails to achieve traceability.

It’s important to acknowledge that challenges from traditional machine learning approaches, such as bias and unfairness, persist. Adhering to responsible AI principles is crucial for the successful implementation of these new models. To ensure ethical and effective use, it’s essential to follow established frameworks for responsible AI development, such as the one outlined in our Responsible AI Framework.

By prioritising responsible AI practices, we can harness the power of generative AI while mitigating potential risks and fostering trust in these transformative technologies.

Pioneering progress: the road ahead for generative AI in insurance

Creating custom, state-of-the-art generative models is currently the domain of specialised companies. Nonetheless, the swift pace of development and frequent research publications are making it increasingly accessible for non-specialised firms to adapt and extend existing models or develop their own models.

While widespread adoption of generative AI in business may still be a few years away, gaining experience with these innovative models is essential to remain competitive in a rapidly evolving landscape.

By embarking on your generative AI journey now and implementing initial use cases, your company can stay at the forefront of this transformative technology. Establishing generative AI flagship projects using non-sensitive data that deliver tangible business value can not only raise awareness within the organisation, but also nurture an AI-co-creation mindset throughout the company.

Check out our dedicated ‘Generative Artificial Intelligence for Business’ training programme to delve deeper into the technical aspects of generative AI, its constraints, and detailed use cases across multiple industries. Or take advantage of our customised workshops for a tailored exploration of potential AI applications across your business, with a focus on  your unique goals and requirements.

Learn more about machine learning technologies and how to optimise and grow your organisation with the right AI solution

Contact person for Switzerland

Markus Reding

Managing Director Insurance Switzerland & Partner

Markus Reding leads the Market Unit Insurance at Zühlke in Switzerland. For more than 20 years he is responsible for innovation, strategy, product management, software engineering, and business development in various leadership positions and has practical experience from numerous digitisation projects. Meeting the challenges and market trends in the insurance industry with innovative solutions is what drives him.

Contact
Thank you for your message.
Contact person for Germany

Gabriele Baierlein

Director Business Development & Partner

Gabriele Baierlein, who joined Zühlke in 2016, is the Director of Business Development & Partnerships for the Zühlke Group. She has many years of experience in cross-industry sales and management, most recently as Market Team Lead at Zühlke, where she oversaw business development and the service portfolio for the consumer goods industry. In her current role, Ms Baierlein is driving the development and expansion of the Financial Services segment with a focus on the insurance industry in Germany. She is also a lecturer in business administration and project management at the University of Applied Sciences Munich (FOM) and the Chamber of Commerce and Industry in Bavaria.

Contact
Thank you for your message.
Contact person for Austria

Helmut Taumberger

Managing Director Cross Markets, Austria

With his years of experience and a strong innovative mindset, Helmut Taumberger is digital transformation personified. As a Managing Director Cross Markets in Austria, he is responsible for steering the company’s strategic orientation and development. A qualified engineer, he has worked in the IT sector since 2003 and has lent his substantial expertise to various international businesses. He likes nothing more than coming up with practical strategies and getting people excited about technological change.

Contact
Thank you for your message.
Contact person for United Kingdom

Brewster Barclay

Business Development Director

Brewster Barclay has a long history developing and selling innovative software and hardware solutions in the electronics and Internet industries, including running a start-up for 6 years. He is dedicated to helping customers create innovative solutions in healthcare and has shown this outside of his Zühlke responsibilities in his frequent mentoring of e-health and medtech startups.

Contact
Thank you for your message.