The transition from generative AI to agentic AI marks the major milestone of 2026. While 2024 and 2025 were spent understanding Large Language Models (LLMs) and creating helpful chatbots, the current year finds corporate giants grappling with a much more complex challenge: delegating actual action and decision-making to autonomous digital systems. At the recent Fortune Brainstorm Tech conference, leaders from Citi, Ford, and Experian shared their strategies for scaling these "agents," placing a critical factor at the center: trust.
From Information to Autonomous Action
AI agents are not limited to answering questions. Instead, they can plan strategies, use tools, and execute tasks—from approving a loan to optimizing a supply chain—without constant human supervision. However, for a bank like Citi or an automaker like Ford, the risk of a mistake is enormous. As highlighted at the conference, the difference between a chatbot giving the wrong information and an agent incorrectly transferring funds is the difference between an awkward moment and a catastrophic financial loss.
Shadman Zafar, CIO of Citi, emphasized that the bank's approach is based on "controlled autonomy." Citi does not allow agents to act in a vacuum. Instead, they have integrated safeguards that require human intervention (human-in-the-loop) for any decision exceeding a specific risk threshold. "Trust is not built with perfection, but with the ability to control," he noted.
Ford and Experian: Data and Transparency
For Ford, the challenge is the complexity of physical systems. Using AI agents in the supply chain requires integrating data from thousands of suppliers in real-time. Ford's strategy focuses on "observability." Using tools like those from Dynatrace, the company can monitor not only the outcome of an AI decision but also the logical path the agent followed to get there. This transparency is essential to convince engineers and production managers to trust autonomous systems.
On the other hand, Experian, which manages sensitive financial data for millions of consumers, sees AI agents as a tool for democratizing financial services. However, the ethical dimension is paramount. Experian applies strict protocols to avoid bias in its algorithms. The company's AI agents undergo continuous "stress testing" to ensure that their creditworthiness decisions remain fair and transparent.
The Challenge of Scaling and the Future
Scaling AI agents is not just a technical issue, but an organizational one. Companies must redesign their structures to accommodate these new "digital employees." This means investing in staff training so they learn to collaborate with AI rather than fear it. The consensus among participants at Fortune Brainstorm Tech was clear: the road to full autonomy will be long and will require gradual steps.
"It's not a revolution that will happen overnight, but an evolution of trust," the speakers noted.
In conclusion, Citi, Ford, and Experian show that success in the era of Agentic AI depends not only on the power of the models but on the robustness of governance frameworks. A company's ability to monitor, explain, and correct the actions of its digital agents will be the competitive advantage of the future. As we move into the second half of 2026, the focus shifts from "what AI can do" to "how we can trust it to do it on its own."
