The Strategic Advisory Team within DAC Beachcroft’s claims division (CSG), has released a detailed thought-leadership report into the impact and potential of AI in the insurance claims process – which reveals ‘a sector embracing change, with eyes wide open’. Here are the details;
The team conducted a series of qualitative one-to-one interviews with senior managers from a diverse group of compensator organisations across the claims market, and then tested the common themes emerging from these interviews with a wider group through an online survey. The report also includes a foreword from the Right Honourable Sir Robert Buckland, KBE, KC, former Lord Chancellor and Secretary of State for Justice and now member of DACB’s Policy Unit.
The aim in conducting this research was to help inform clients’ strategic thinking about AI adoption in the months and years ahead. The resulting report provides a review of the claims market’s current and planned use of AI, looking thematically at the impact of new technologies on important stakeholders in the process, such as colleagues and – most importantly – customers.
A number of clear themes emerged, as well as a definite sense of optimism – indeed excitement – regarding future use cases, and what AI can do to drive process efficiency, enhance job satisfaction and improve customer journeys:
1. Whilst many insurers are already using ‘traditional’ predictive AI and machine learning to help assess loss more accurately and identify patterns that point to potential exaggeration or fraud, Generative AI is seen as another level again. The insurers we spoke to are all now piloting use cases to assess the value this new iteration of AI could potentially add.
2. All interviewees talked about the importance of making sure humans remain in the loop and of perfecting the interplay between the roles of the humans and the AI. A number underlined the need for humans to retain responsibility for decisions, AI being deployed to assist this decision-making process, by providing support to claims handlers in a number of ways, freeing up their time to focus on the more complex and human elements of their jobs.
Hardly anyone interviewed for this study anticipated a future where the claims process would be fully automated. Nearly all described human interaction as essential, for the simple reason that an insurance claim is always a time of stress and emotion for customers. This is the reason why insurers’ focus for AI investment, and now Generative AI, is for now firmly on back-office functions, working behind the scenes to enable them to service customers better, rather than on the customer-facing part of their business.
3. Training was mentioned by all interviewees as another key to the successful integration of AI in the claims process, helping the humans in the process to understand not just what AI can do and how to use it, but crucially its limitations too, particularly regarding its interpretation of data.
As a number of interviewees put it, if the data input isn’t perfect, (and given the industry is currently relying on ‘legacy data’, this is most of the time), the people working with these outputs need to understand that. A number talked about significant investments their companies have made in establishing ‘Data Academies’, to educate the people in the business about how to interpret the output from Generative AI and what to watch out for in reviewing that output: anticipating assumptions the AI might be making when processing data and identifying patterns and trends; watching out for bias that could creep in to the process because machines don’t understand the nuances that may sit behind some numerical trends in the way that humans can; and being alert to the possibility of machine hallucinations.
Everyone talked about the critical role for the humans in the claims process, in reviewing and challenging AI output and making sure the conclusions drawn from it make sense in the real world.
4. Most described the major potential benefit of Generative AI being its ability to ‘structure unstructured data’. As one interviewee said, ‘the moment you can do this, you are able not only to make significant improvements to existing models, but also create new models which were out of reach before’.
The three areas of insurance business and claims handling that could benefit most from Generative AI in this way were listed as:
-supporting claims handlers’ calls with customers Generative AI has a lot to offer in terms of creating summaries of case information for claims handlers at great speed and producing transcripts of calls with customers, freeing them up to focus on the more complex, human and interesting parts of their job.
-processing the millions of incoming documents that insurers receive each year, having the ability to extract the most useful and relevant information much more quickly.
-verifying images Increasingly customers’ claims are supported with images of damage, and in these days of AI image generation, it is easier for the unscrupulous to fabricate these. Generative AI is able to assess images submitted in a case to check where they have come from, whether they are genuine, or whether there is fraud at play.
DACB’s Strategic Advisory Team provides insights and advice on a broad range of nascent issues and innovations impacting general insurers. With the rapid rise of technologies like Generative AI – and now the emerging shift towards Agentic AI – and the continuing pace of change, the claims process could look dramatically different in just five years. The idea behind this report is to help the insurance industry navigate this uncharted territory.
In launching the report, Pete Allchorne, partner at DACB and Head of the Strategic Advisory Team, said:
“AI is a hot topic just now, but there is a lot of hyperbole and PR noise around it, which confuses the picture for those needing to make proper and diligent assessments to guide their organisations into the future. This report is designed to bring some much-needed clarity.
“Many organisations already use rules-based RPA (Robotic Process Automation) for processes involving repetitive tasks such as data entry and extraction. The opportunity of AI is to take this to an entirely different level: AI promises the ability to identify and learn from patterns in data, allowing systems-driven decision making, with or without a ‘human in the loop’, in order to solve complex problems, which previously required human thinking. Our aim was to chart the insurance industry’s adoption so far, and its future plans.
“We hope this report provides food for thought for the industry and that it helps advance conversations within insurance organisations and teams.”
Sir Robert Buckland said:
“This report is fascinating reading for anyone who is interested in the impact of machine learning. It is a timely contribution as the UK faces some key decisions about the regulation of AI use in the public and private sectors. Legislation on AI is yet to emerge, and its likely content and approach remains unclear as the Government continues to consult over the next year or so.
“Across both one-to-one interviews and survey responses, a clear picture emerges from this report: the insurance industry is embracing the potential of Generative AI with enthusiasm, tempered by a thoughtful and measured approach. This is not an industry rushing blindly into the future – it is one recognising both the opportunities and responsibilities that come with technological change.”
You can view the Report in full here: https://csg.dacbeachcroft.com/ai-in-claims/ai-in-claims-title
Be the first to comment