Insurers, intermediaries and UMAs stride into the AI age
Few things make your writer happier than a sub-45-minute panel discussion, and here are three reasons why. First, the shorter the session, the less background ‘fluff’ there is to churn through in the write-up. Second, it allows for a sharper take on different views around a common theme. And third, it eliminates some of the pain in repeating the same name ad nauseam. So, no more Jane said; according to Jane; Jane explained; in Jane’s view; here’s what Jane had to say about it, etc.
The ‘What is AI’ opener
The closing presentation to the SAUMA Conference 2025 lived up to your writer’s panel discussion expectations by offering three experienced non-life insurer views on the future of underwriting in an artificial intelligence (AI) setting. Santam Insurance’s Manager: Emerging Brokers, Letlhogonolo Tau, threw out an unexpected opener, asking each participant to explain what they understood by the term AI.
Philippa Wild, Chief Underwriting Officer at Bryte, rose to the challenge. She explained AI as a machine or computer capable of gathering and interrogating data in a predictive way and responding to user inputs in a way that makes sense. Deen Pillay, Portfolio Manager at Emerald Underwriting Managers, offered a handy AI = AI formula to frame his response. He held up automated insurance as a proxy for AI. “Insurance brokers and underwriting management agencies (UMAs) can use AI-backed tools to make their lives easier,” he said. The caveat: AI is not a silver bullet; it merely enhances a firm’s capabilities and efficiencies.
Last up, Frans Prinsloo, in charge of business development and innovation at Lombard Insurance, said, “From an insurer perspective, the AI model that is most commonly being used centres around machine learning.” He predicted that AI-backed insurance automation would evolve, with generative AI bringing choice, insight and prediction to underwriting decision-making, eventually even handling certain business processes outright. “AI is going to play an important role in our lives, and it is something that we need to embrace and understand in underwriting moving forwards,” he said.
Integrating technology into insurer functions
Tau challenged the panellists to offer practical examples of how AI was being integrated into insurer functions. There was consensus that the ability to use live data from multiple sources to inform pricing decisions was a standout AI-linked underwriting development. So, for example, an insurer’s pricing model can reference up-to-date claims information at quote stage. “The decision remains with the [human] underwriter, but AI enhances underwriting and provides insights [informed by both] internal and external information,” Prinsloo said.
Early adopters of emerging technologies will have to walk a fine line to maintain trust across the value chain. The panel suggested that clear delineations be made between human and machine, especially as AI and generative AI find traction across insurer, intermediary and UMA functions.
Wild singled out improved client servicing outcomes, lower costs and process efficiencies as potential spin-offs from AI adoption. She shared a personal experience of the time saved by using a generative AI like Microsoft Copilot to prioritise an email inbox after returning from annual leave. “The AI tool frees up my time to put my mind to other things,” she said. From an underwriting perspective, any tool that streamlines data-intensive research and comparison frees up staff to get the final underwriting decision right.
“AI has a role to play in enhancing knowledge, enhancing decision-making and providing insights that help the underwriter,” Prinsloo added, stating that AI would play a role in augmenting non-underwriting tasks, leaving the human underwriter as a key part in the process. Pillay, meanwhile, urged attendees to take a client-centric approach to tech innovation.
“It’s useless spending money on AI if it does not do anything for your clients,” he said. Returning to the AI = augmented insurance framing, he noted that AI was there to streamline processes and introduce more excitement to the human side of underwriting. One of the panellists seized on this light-hearted interlude by declaring that many insurers and UMAs were employing more data scientists than actuaries. On a more serious note, insurers were to think beyond providing capacity and a licence to write on, to assisting UMAs with their data culture.
The biggest differentiator in the AI age: Data
Data and a firm’s ability to gather, store and utilise it will be a huge differentiator as the insurance sector pushes into the AI realm. “Data is a challenge,” said Prinsloo. “We must figure out how to use technology to get our data to a place where we can use it effectively to make good decisions.” AI stands out as the tool that will help brokers, insurers and UMAs to leverage data (from both internal and external sources) to achieve a sensible balance between exposure, risk and price.
Leveraging data and streamlining manual processes featured throughout the panel discussion as areas in which AI tools could reinvent underwriting. According to Pillay, 80% of an underwriter’s time is spent on admin work. He envisaged a future world where, upon receiving a quote request, various AI agents would instantaneously start working on sales-related admin alongside the all-important risk assessment, risk management and pricing functions. “AI is an innovative tool that will take our businesses to the next level,” he said.
Tau asked the panellists how AI and automation might impact on human relationships. Wild held that new technologies freed up time for staff to focus on relationships and service. Machines can complete complex assessments and comparisons, creating huge efficiency gains and freeing up time for team members to work on broker-client, broker-UMA, and UMA-insurer relationships. The consensus was that AI would never replace relationships, but rather enable underwriters to select risk more efficiently.
AI adoption risks
But there are risks to fast-tracking technology too. According to Prinsloo, the biggest risk stems from allowing unchecked AI deployment. “AI can generate biases that can influence underwriting negatively, and [you will have to] check on the results coming out of AI,” he said. “You cannot allow the machine to take over and create biases that risk changing the landscape.”
Pillay concurred, saying that AI adoption could create new risk pools. He identified the rapid roll-out of AI in facilitating cybercrime worldwide as one example. To paraphrase his comment: AI is not a system, but a set of tools that enable your systems to do whatever you tell them to do. You have to cater for emerging risks in the context of how AI utilises data. He believes there will be a “coming together” of insurtech start-ups, insurers, and regulators to ensure that AI-related data use does not overstep moral boundaries.
Pillay also commented on the value of an insurance brand, saying that a track record in doing insurance would hold more sway than a technology label among future consumers. “People buy reputation and brand; we are the experts in insurance, and our technology partners enable us to do what we already do better, i.e. selling good quality risk at the best possible prices,” he said. The discussion ended with the familiar ‘last comment’ round robin.
Let generative AI be your pilot
Wild recommended that brokers, insurers and UMAs use AI tools to guide their AI adoption. You can literally use a free large language model like ChatGPT to flesh out your approach. Pillay urged attendees to take the first step on their AI journeys by reviewing processes and making small iterations along the way. “AI is good at managing volumes; it does so 24 hours a day, 365 days per year without tiring,” he said.
And with the last word, Prinsloo said, “Data science as a culture should be part of the fabric in your business; you should create that culture as a community body and guide it from there.”
Writer’s thoughts:
As insurers charge ahead with AI adoption, intermediaries, UMAs and other third-party providers will have to adapt or risk falling behind. Is your firm considering any changes in response to AI-enabled underwriting decision-making? Please comment below, interact with us on X at @fanews_online or email us your thoughts [email protected].
