Sid (Siddharth) Dixit is a principal architect at CopperPoint Insurance Companies with over 15 years of expertise in digital transformation.
As we move into the end of 2025, the insurance sector is at a pivotal moment, fueled by rising climate losses, narrow margins and fast-rising maturity in artificial intelligence (AI). What was once an experimental technology is now a mainstream imperative to survive and thrive.
Generative AI in insurance will grow from $1.09 billion in 2025 to $14.30 billion in 2034. With almost 90% of insurers now deploying AI, leaders are integrating it across three strategic pillars: preventive risk management on a proactive basis, hyper-efficient operations and resilient digital foundations, all of which rest upon a human-in-the-loop approach to keep the AI safe, ethical and productive.
From Prediction And Repair To Prediction And Prevention
By far the greatest influence of AI is a shift from a reactive claims settlement model to a proactive services model motivated to avoid losses. Insurers become positive collaborators in the avoidance of losses and employ technology such as Internet of Things (IoT) and video analytics powered by AI. But AI does not act in a vacuum—the function of human judgment is to spot sophisticated risks, authenticate findings and step in where algorithms fail.
For example, on commercial lines, Intenseye’s computer vision-based video analytical systems keep a watchful eye on workplace hazards such as non-compliance with personal protective equipment (PPE) use. Such software, though highly successful, works best if frontliners and safety managers remain apprised about preliminary results reached by AIs, tweaking alerts ideally and supporting a safety culture. Human judgment averts AIs’ interventions from having a disruptive or even detrimental effect.
Rewiring Core Processes To Be Hyper-Efficient
AI is modernizing the backbone of insurance: claims and underwriting. In claims, multimodal AI can ingest text, image and audio all at once from a first notice of loss (FNOL) to facilitate smart triage, fast-tracking and improved fraud detection. But it’s not only automation that’s shaking up the business status quo; it’s collaboration between AI and veteran claims experts.
U.K. insurer Aviva, for instance, implemented over 80 AI models, reducing assessment time in difficult cases by 23 days and saving over $82 million in 2024. But accuracy and customer trust were dependent on retaining claims adjusters in the equation for edge cases and high-stakes decisions.
In underwriting, AI automates data extraction from messy, unstructured submissions, letting underwriters focus on complex judgment calls, but automation alone isn’t enough. What is truly needed is augmentation. Human-in-the-loop workflows that involve actuaries allow underwriters to verify, override or augment AI recommendations, ensuring nuance and expertise are preserved.
Chatbots and automated quoting systems may handle routine questions, but complex negotiations still demand experienced human insight. Similarly, anomaly detection tools flag operational risks, but humans must interpret those signals and decide on corrective actions.
Constructing Pillars For A World Tomorrow
Releasing the full potential of AI exceeds algorithms. It requires foundational investments in people-oriented AI, starting with leading-edge data infrastructures, governance frameworks and, most importantly, talent.
Insurers still held back by legacy technology cannot deliver the clean, consistent data required by AI. Even if technology is available, low-quality data can confound models until humans actively authenticate, correct and train these systems. Autonomous-learning AI can only remain relevant and accurate with ongoing feedback from domain experts.
As applications of AI continue to multiply, regulatory attention continues to mount. Successful governance is not an option—it’s survival. To address mounting standards around fairness and transparency, insurers must employ effective checks to check for bias and offer explanations behind decisions made in AI. Major insurance companies are developing “trustworthy AI” frameworks, but these only remain successful if operations are continuously checking and interpreting outputs of AI.
A large skills shortage could jeopardize AI deployment. Although 92% of insurance professionals desire generative AI competencies, only 4% of insurers are substantively reskilling their personnel on a large scale. That’s unsustainable. Insurers increasingly recognize the importance of upskilling employees to ensure AI tools enhance human expertise rather than replace it. But this only happens if people can be trained, trusted and brought into loops.
The future of insurance is not AI versus human. It’s AI with human. The most forward-thinking insurers are those treating human-in-the-loop not as a safeguard but as a strategic enabler. From risk prevention to claims automation and ethical governance, embedding human expertise at every step is the key to making AI smart, safe and scalable.
In personal insurance, for example, when a customer calls after losing a spouse or their home, no script or algorithm can fill that silence. They want speed, yes, but they also want empathy. They want to be heard. That’s why the human-in-the-loop isn’t just a technical framework—it’s a moral and emotional imperative. It ensures that even in a world increasingly shaped by algorithms, people still come first.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
