Igloo’s transformation into an AI-native insurtech is not just about deploying algorithms. It is a fundamental reengineering of the company’s operating model—one that spans consumers, agents, claims teams, finance departments, and marketing units. But as Co-Founder and CEO Raunak Mehta emphasizes, automation is meaningful only when paired with rigorous human oversight, regulatory alignment, and market sensitivity.
“The premise of using agents is that they have to be autonomous in nature,” Mehta says. “But before they become autonomous, there has to be a certain degree of evaluation and observability. Only when they achieve that do they operate independently.”
This balance is especially critical in claims, where speed and accuracy define customer experience. Igloo’s AI deployment has cut first-response times from one to three days down to just minutes. “Given the last six to seven years of knowledge base we were holding, our teams trained the models to give instantaneous feedback,” Mehta explains.
“Whether documents are sufficient or whether a claim is invalid—that all happens instantly.” For certain Indonesian partners, Igloo’s AI even manages claims end-to-end, including adjudication. Yet Mehta stresses that not all agents are autonomous: “Quite a few still require Human-in-the-Loop. Humans play a very important role in bringing AI to a level where it starts operating on its own.”
Fraud detection introduces new complexity. As generative AI tools grow more accessible, the ease of producing synthetic evidence increases. “With people generating real images or evidence via AI, it is more of a problem than before,” Mehta says. Igloo trains its agents on product-specific fraud patterns and leverages its accumulated data to detect anomalies. But Mehta acknowledges that this will become one of the biggest challenges across financial services. “This is going to be a problem for the overall industry,” he cautions.
One of Igloo’s most innovative use cases lies in agent support. Field agents across Southeast Asia often resist downloading dedicated insurance apps, leading to uneven adoption. Igloo countered this by integrating its AI tools directly into messenger platforms. “Most people are already proficient on Viber, WhatsApp, or Zalo,” Mehta says. “If we can provide them a conversational way to get quotations and close sales, we break through the adoption hurdles.” The AI Sales Admin Chatbot enables agents to retrieve quotations instantly through natural-language prompts, dramatically shortening sales cycles.
In finance and marketing, automation is creating new efficiencies. Financial reconciliation, which Mehta describes as “a repetitive, complex exercise,” has already been optimized using AI. In marketing, Igloo envisions a full chain of AI agents—from campaign ideation to execution and targeting. “If we can deploy agentic AI solutions that create campaigns, create a season calendar, execute them, and then use workflows to target the right customers, marketing can run 24/7, 365,” he says.
Operating across six Southeast Asian markets requires strict alignment with data residency and regulatory frameworks. “In insurance, data needs to be within the country,” Mehta notes. “Our deployments are in line with compliance requirements in each market. AI just happens to be part of it.” For partners with legacy systems, Igloo deploys smaller, quantized models that can run on-premise. “You don’t need massive expensive systems for enterprise use cases,” he explains. “You can work with very small models and get good cost and latency benefits.”
Despite Igloo’s strong automation capabilities—including 98 percent adjudication accuracy and 70 percent reduction in manual claims work—Mehta insists on maintaining a strong governance layer. “AI agents also suffer from efficiency decline if you don’t upscale them,” he says. “Because they are not deterministic. Ask an LLM a question ten times, it might give different answers. All this needs to be managed.”
Looking ahead, Igloo expects AI to contribute significantly to its goal of breaking even in 2026 by improving margins, raising conversions, and enhancing engineering output. “Usage of AI leads to increased sales because conversion improves,” he says. “And we don’t have to expand our operations team linearly. But of course, we will need to make the right capital expenditure.”
Mehta believes the ultimate frontier lies in AI-driven product innovation. Igloo is now training models on insurance products across the region to generate new, untapped product designs. “A lot of insurance products today are incremental variations,” he points out. “If you provide AI with the right grounding, it can give you products that can solve the penetration problem faced by most countries.”
For an industry defined by caution and tradition, Igloo’s AI-native playbook is unconventional, ambitious, and increasingly influential across Southeast Asia. As Mehta puts it, “We are maturing as a company, and almost everything is going to have some flavor of AI. This is the future we are building toward.”