Kathmandu. Insurance companies are beginning to use agent artificial intelligence (AI) to automate underwriting, document verification, and customer service. However, human oversight is still required for financial, underwriting, and regulatory risk decisions.
Speaking at the Asian Banking & Finance & Insurance Asia Summit in Singapore, Gonzalo Carvalho, APAC Director at German IT consulting and software solutions firm Valentique, said insurance companies are moving away from AI assistants to AI agents. “Especially in underwriting workflows, it allows it to independently execute parts of the business process,” he said.
According to Carvalho, insurance companies need to build governance in these systems to reduce misinformation and maintain auditing as AI adoption accelerates. “AI is something you know,” he said. You need to use it in the right context and railing. Which is very important. However, it is also difficult to structure it properly. ’
Enterprises are expected to see a significant increase in AI projects in the next few years, Carvalho said. “While many tech leaders remain cautious,” he said, “by 2028, we expect to see a 2,500% increase in AI-related projects.” In addition, 67% of IT leaders still don’t feel comfortable with AI. ’
According to Carvalho, AI-assisted software development is also becoming the standard. “90% of engineers are expected to use AI-assisted development tools and 55% of developers are expected to use multiple coding tools at once,” he said.
Carvalho said rising productivity also created new governance challenges. “With AI, development is becoming much easier. But the key question is, are we building secure, scalable, and properly governed solutions?”
To demonstrate how agent AI is being used in insurance, Carvalho demonstrated a life insurance workflow where AI agents verify a customer’s identity, proof of address, and medical information before submitting an application for insurance.
AI also generated risk assessments and allowed insurance companies to ask about the rationale behind their recommendations via a chatbot interface. During the demonstration, the AI assessed the applicant’s risk as 68%. Carvalho, who acted as a human underwriter, overrode the recommendation and approved the application. “Because I’m a human being and I have the ability and the right to do it. So I will decide to approve,” he said. But it’s my ability to approve. ’
The workflow also generated offer letters, customer e-mails, audit logs and recorded whether the work was performed by AI agents or human users. “Agentic AI can automate repetitive insurance tasks. But organizations still need clear governance about how AI systems verify information, generate recommendations, and record decisions,” Carvalho said. The ultimate responsibility for underwriting decisions rests with human reviewers. —Agency
