Most of the conversations about Amnesty International agent Focus on the extent of independence of these systems. Financial services leaders are asking a different question: Where should autonomy stop?
For banks and insurers, the challenge is not just deployment Amnesty International Agents. It is determining how those systems, governance requirements, and human expertise can work together in decisions that have financial consequences.
During a recent discussion among leaders from JPMorgan Chase, Allianz, AWS, and SAS, one theme came up again and again: The future of agentic AI is not about removing humans from decisions. It’s about redefining their role.
From human in the loop to human in the lead
For years, debates have been around Artificial Intelligence Governance Focused on keeping people “in the loop”. But some organizations have begun to look at this relationship differently.
As Andrea Polman of Allianz said:
“They stopped using the human in the loop. It’s the human in the front.”
This distinction may seem subtle, but it reflects a tangible shift in how financial institutions approach AI adoption.
Instead of asking where AI can replace people, organizations are evaluating where AI can support them. Agent AI can gather information, summarize findings, identify patterns, and recommend next actions, but responsibility for important findings remains with humans.
This approach reflects the reality of banking services insuranceDecisions often carry financial, regulatory and reputational consequences. While AI can help speed up operations, organizations still need people to assess risks, exercise judgment, and maintain trust with customers.
Where agentic AI creates value today
The strongest opportunities for agentic AI are not fully autonomous systems. Instead, many organizations find value in repetitive, time-consuming, and highly manual operational processes.
At JPMorgan Chase, Adolfo Lopez described many current AI applications as “assisted or delegated,” helping employees enhance processes and support decision-making rather than acting autonomously.
As a result, you see many organizations KYC and customer onboarding As practical areas to support agentic AI. Artificial intelligence agents It can help collect information, identify missing documents, and automate parts of the workflow, allowing employees to spend less time on administrative work and more time serving customers.
As AWS’ Sri Raghavan points out, “A lot of manual tasks are performed as part of the KYC process.”
Insurance organizations are See similar opportunities In processing claims. Agentic AI can help collect information, review documents, and route cases more efficiently, allowing claims professionals to focus on more complex customer interactions and decisions. Allianz has already seen early success using AI in high-volume claims scenarios where speed and efficiency are critical. In these cases, AI can speed up the process, but humans remain responsible for the final decision.
Across these use cases, the goal is not to remove human involvement. It is to reduce the administrative burden so that employees can focus on higher value work.
Why is human oversight still important?
If these use cases prove their value, why aren’t organizations moving more quickly toward autonomous decision-making?
The challenge becomes more pronounced in high-stakes decisions. Think of a hypothetical underwriting workflow. An AI agent can collect the applicant’s information and reference supporting documents and create a recommended risk assessment. But because lending decisions have significant financial and regulatory implications, organizations still need clear oversight, explainability and accountability for the bottom line.
According to participants, technology may not be the biggest obstacle.
As Lopez explained:
“The technology is there. But it’s the regulations that limit us.”
Financial institutions operate in some of the most regulated environments in the world. Activities such as lending, underwriting, claims processing and risk management are subject to strict requirements of fairness, transparency and accountability. An incorrect decision can lead to legal, financial and reputational consequences.
These realities are shaping how organizations deploy agentic AI. While many organizations are comfortable using AI to gather information, analyze data, and support decisions, they remain cautious about allowing AI systems to make consequential choices independently.
Polman pointed to insurance as an example. While AI can aggregate information from many sources, it emphasized the importance of maintaining human responsibility for risk-related decisions, especially in complex cases.
In other words, the challenge is not the ability of AI to contribute to decision-making. It decides decisions that still require human judgment and accountability.
A new operating model for agentic AI
Organizations that succeed in using agentic AI may not be those that seek to achieve the highest level of autonomy. These companies may be the ones that most effectively combine intelligent automation with human expertise, governance and accountability.
The question is not whether humans will still participate. And this is how their role is evolving as trust in AI systems grows.
As organizations continue to explore agentic AI, the most effective operating model may be the one described by Polman: not a human in the loop, but a human at the forefront.
In financial services, the question is not how much autonomy AI can achieve. It is where autonomy creates value and where human judgment must remain.
