Spare a thought for claims handlers. Day after day, they face a myriad of challenges from sifting through lengthy, intricate documents to extracting crucial information while ensuring nothing is overlooked. They must manage data with care, ensure information is not lost through multiple touchpoints, stay compliant and keep their stakeholders satisfied.

The complexity and time-consuming nature of this task have traditionally demanded extensive manual effort and expertise. However, the advent of AI technology is set to revolutionise the landscape, introducing tools that can assist claims handlers by searching documents, extracting information, writing summaries and suggesting responses to stakeholders.

Claims handling AI is part of a wider trend where we can soon expect to see an AI assistant embedded in every
workplace tool. Microsoft, for example, has very recently launched CoPilot, an embedded AI assistant for Microsoft Teams, Outlook, Word, Excel, and PowerPoint. Embedding AI assistants in claims handling tools is the logical next step.

Must-have attributes

Drawing on our own journey in creating Noria’s “ClaimsBuddy” AI Assistant, w’ed like to share some of the essential attributes for marine insurers to look for when building or selecting an AI claims handling assistant. First, an effective AI claims handling assistant should excel in efficiently parsing through extensive documents, extracting essential information and offering summaries. Look for a tool that simplifies and expedites the process, minimising the time spent on reading and analysing complex data. For example, the assistant should be able to search a lengthy document to create a one-page summary of a ship’s maintenance history. Early data indicates our ClaimsBuddy creates efficiency savings between 25% to 50%, benefiting both the insurer and the customer.

Humans still in charge

Second, human oversight and collaboration remains key. While AI greatly enhances efficiency, its role at present should be to function as an assistant, not a replacement, and should require human validation and review. We would caution against engaging with any software provider that asserts that their claims handling solution can
completely automate claims handling, as mistakes are expensive to correct.

Ease of use is critical

Third, the user experience of the AI assistant should be intuitive, requiring minimal to zero training. However, we would not suggest that easy UX means any layperson can use a tool built for such a specific use-case. For now, AI claims handling assistants still need skilled, knowledgeable and experienced human oversight. Ease of use will convince even the most traditionally-minded claims handler to adopt the tool rather than continuing with inefficient ways of working. As more companies adopt AI assistants, candidates are likely to gravitate towards the AI-haves at the expense of the have-nots.

Training is all

Fourth, the right training is essential. Ensuring that the system has undergone extensive and industry-specific
training greatly influences its capability to accurately extract pertinent information. A well-trained AI assistant, built on a foundation of comprehensive and relevant data, ensures its precision in understanding and processing the intricate details within marine insurance claims. The depth of training impacts the system’s ability to deliver accurate, efficient, and industry-specific support to claims handling professionals.

Constant learning and improvement is also key. An adaptable assistant should continuously improve with additional training and input, becoming more precise and accurate in providing responses and extracting crucial data. With the right human oversight, this learning capability will enhance the system’s accuracy over time.

Holistic support

Holistic support and assistance is also important. An ideal assistant goes beyond mere extraction of information. It should assist in providing responses, recommendations and valuable insights to aid decision support. Industry relevance and specificity is another area to be taken into account.

For the marine insurance industry, the AI claims assistant should be capable of understanding industry-specific
context, language, nuances and challenges. Look for a tool tailored to the specific needs and intricacies of the marine insurance sector and compare its outputs with that of a generalised model such as ChatGPT. A clear path to evolution and integration must also be taken. An AI assistant should provide a clear path for its
integration with existing claims handling systems. It should also showcase potential for evolution and adaptation, integrating new functionalities and updates seamlessly.

Look into the future

And, future compatibility and use cases must be taken into account. Talent recruiters tell us that the best human hires score high in adaptability and willingness to learn. We would suggest the same criteria should be applied to any AI tool.

Seek an AI assistant that is capable of adapting to future trends in claims handling. Look for tools that promise future integration and extension to other areas, enabling adaptability to evolving needs and technologies. “Talent recruiters tell us that the best human hires score high in adaptability and willingness to learn.

We’d suggest the same criteria should be applied to any AI tool. Seek an AI assistant that is capable of adapting to future trends in claims handling.” As for future use cases in banking and insurance, we can expect AI support across customer communication, credit processes, decision support, advanced automation, sales support, as well as document control and verification.

We would stress again that the AI tool should not replace but complement human claims handling expertise, serving as an aid rather than an automation solution. Assistive AI improves quality and efficiency and lets you use your human skills where they matter most. By choosing a solution with these attributes, the right AI claims handling assistant has the potential to become an indispensable ally, enhancing efficiency, accuracy, and
overall quality, while upholding the expertise and oversight that only a human touch can provide.