Insurers are beginning to deploy agentic artificial intelligence to automate underwriting, document verification, and customer servicing, but human oversight remains essential for decisions that carry financial and regulatory risk, according to Gonçalo Carvalho, APAC Director for German IT consultancy valantic. Speaking at the Asian Banking & Finance and Insurance Asia Summit in Singapore, Carvalho said the industry is moving beyond basic AI assistants toward systems that can execute parts of business processes independently.
From AI assistants to autonomous agents
The shift is happening fastest in underwriting workflows. Rather than simply suggesting next steps, AI agents now verify customer identification, proof of address, and medical records before routing applications for human review. Carvalho demonstrated a life insurance workflow where the AI checked these documents and generated a risk assessment. When the system flagged an applicant's risk at 68%, he overrode the recommendation and approved it himself. "Because I'm a human and I have the capacity, authority to do it, I will decide to approve," he said. "It's not on your capacity but on my capacity to approve."
The workflow also produced offer letters, customer emails, and audit logs that clearly recorded whether actions were taken by AI agents or human users. Carvalho said that while agentic AI can take over repetitive verification tasks, final accountability for underwriting decisions stays with people. This combination of AI Agents & Automation and human judgment is changing how carriers balance speed with compliance.
Governance becomes a hard requirement
The move toward autonomous AI agents creates new governance pressures. Carvalho noted that enterprises expect a 2,500% increase in AI projects by 2028, yet "67% of IT leaders don't feel comfortable with AI at the moment." He pointed to the rapid spread of AI-assisted software development-where 90% of engineers are expected to use such tools and 55% will run multiple coding tools simultaneously-as a sign of both productivity gains and emerging vulnerabilities. "With AI, development is becoming much easier. But the important question is: are we building solutions that are safe, secure, scalable, and governed properly?" he asked.
For carriers adopting AI for Insurance, the challenge is embedding guardrails that reduce hallucinations and maintain a clear audit trail. "AI, you need to use, as you know, the right context and guardrails, which is very important but also difficult to structure properly," Carvalho said. Without that structure, he warned, organisations risk building systems that produce outputs they cannot fully explain to regulators or customers.
Why this matters for insurance professionals
Agentic AI will increasingly handle the verification and analysis that slows down underwriting and claims processes. But the demonstration made clear that the technology alone isn't making final calls-humans are. Underwriters and claims managers who know how to interrogate an AI's reasoning, override its recommendations when needed, and document those decisions will be the ones keeping the business both efficient and compliant. The tools are changing; the need for professional judgment is not.
Your membership also unlocks:
