The Fourth Industrial Revolution has been advancing more slowly than expected, but generative AI is accelerating change, says Justin Davies, head of region, EMEA, Xceedance.
As the titans of global business and economics gathered at the World Economic Forum in Davos in January, attention focused on the theme of ‘rebuilding trust’. But eight years ago, the hot topic was the Fourth Industrial Revolution (4IR), which was also the subject of a book by WEF founder Professor Claus Schwab.
He stated that 4IR would follow on from the three earlier revolutions driven variously by steam, electricity and computing. 4IR would be a confluence of physical, cyber and biological networks involving technologies such as telematics, Robotic Process Automation (RPA), augmented reality, machine learning, blockchain, the internet of things and drones.
Predictions included machines talking to machines without the need for human input, with virtual assistants such as Siri and Alexa cited as the forerunners of this technological revolution.
In 2017, I wrote about the potential impact of 4IR on insurance for this publication. It’s now time to analyse where we are in this much anticipated revolution, and what the latest advances mean for our sector.
While some predictions, such as the use of drones to assess property damage instead of a human engineer or medical implants that supply real-time health information, have yet to come to fruition at scale, generative artificial intelligence (GenAI) has great potential to improve the efficiency, accuracy and profitability of many aspects of the insurance industry.
The 2016 conference in Davos predicted that 65% of primary school children would be employed in jobs that did not then exist.
Several new jobs have indeed emerged or evolved significantly since then due to advancements in technology, changes in consumer behavior, and shifts in the global economy, including: data scientist; machine learning engineer; augmented reality developer; cybersecurity analyst; drone operator; virtual reality developer; chief diversity officer; and podcast producer.
One role that certainly did not exist in 2016, but which I believe will be of vital importance to the insurance industry, is that of the generative AI prompt engineer. This is a person who can ask the GenAI app the right questions to extract an appropriate response. As with many things, the answer you get is only as good as the question you ask – so if insurance businesses want to get the best of these technologies they will need the people using them to have insurance expertise so that they can ask the right questions.
Prompt engineers are currently in high demand as companies explore ways to effectively use AI tools. This new role is part of a significant increase in demand for employees who understand and can work with AI. The number of LinkedIn posts referring to generative AI rose 36-fold between 2022 and 2023, and the number of job posts containing ‘GPT’ increased by 51% between 2021 and 2022, with some not even requiring a tech background.
With increasing numbers of insurance companies anxious to find out more about generative AI, Xceedence recently demonstrated GenAI’s transformative power in revolutionizing risk assessment and underwriting. An extensive property survey report was uploaded and analyzed using the ChatGPT platform integrated into the secure Azure OpenAI Service - with remarkable results. Traditionally, an underwriter can spend hours reading such reports, but with AI assistance, we showed that GenAI could almost instantaneously extract key highlights, ensuring no critical information was overlooked, and was able to comprehend the report with unprecedented speed.
The study examined the use of generative AI tools throughout the underwriting lifecycle. Results were promising: we found many ways in which GenAI could reduce costs and improve customer experience, business insights and operational efficiency.
At the Xceedance Generative AI Center of Excellence sandbox, we are experimenting with high-ROI generative AI use cases, which include:
•Real-time comparison of various product and coverage options.
•Submission ingestion and processing.
•Submission triaging and prioritisation.
•Risk assessment and rating.
Other applications may include mapping data from claims adjustment reports; extracting the legal entities from multiple documents during sanction checking; and submission data extraction.
We also demonstrated the use of generative AI in policy data extraction and quality assurance, data summarization and extraction from risk engineering reports, demand letter and broker emails, and policy admin and/or product configuration support via a smart ChatBot.
Over the last year, generative AI has attracted a great deal of interest and investigative buy-in from insurance industry senior managers, largely down to the increasing popularity and take-up of ChatGPT.
While more progressive insurers are better positioned to adopt GenAI solutions and plug-ins, firms using more traditional processes can also make use of new technology to create efficiencies. However, companies should proceed with caution as our industry holds an extensive store of sensitive, personal data. It’s advisable to train the large language models used in generative AI in the private sphere to prevent data leakage. It’s also wise to bear in mind that generative AI is trained on internet data, which can be subject to racial, gender, social and ageist bias.
While the pace of technological change has not evolved as anticipated in 2016 at Davos, the increasing prevalence and use of generative AI could be the catalyst that truly accelerates the move into the fourth industrial revolution.
However, technology on its own won’t transform the insurance industry. Generative AI can never replace decisions based on an organisation’s experience, values and culture. The output will only ever be as good as the prompts that are put in. Only by marrying technology with insurance expertise can real change be delivered for the industry.