The global financial sector stands at a critical technological crossroads. For decades, the user experience in mobile banking has been defined by rigid interfaces—a labyrinth of menus, sub-menus, and buttons that demand the user possess a technical understanding of the bank’s architecture to perform even the simplest tasks. However, a seismic shift is underway. Credicorp, the financial powerhouse behind the Banco de Crédito del Perú (BCP), is spearheading a transition toward "agentic" artificial intelligence, aiming to replace static digital navigation with fluid, natural-language conversations that don’t just inform, but actively execute financial commands.
Main Facts: The Transition to Agentic Banking
The core of this transformation lies in the evolution from traditional chatbots to AI agents. While a standard chatbot acts as a repository for frequently asked questions, an AI agent is designed to understand intent, navigate complex logic, and interact with backend systems to complete processes.
Under this new paradigm, a customer will no longer need to navigate through a "Transfer," then "Select Recipient," then "Confirm Amount" workflow. Instead, they might simply type or dictate, "Send $200 to my daughter for her books," and the AI will handle the authorization, verification, and execution. This transition from "navigational banking" to "conversational banking" represents the most significant change in retail finance since the introduction of the mobile banking app itself.
The Chronology of Digital Evolution
To understand where Credicorp is headed, one must examine the progression of their digital strategy:
- Phase 1: Digital Migration (The Foundation): The initial focus was simply to get customers out of physical branches and onto mobile platforms, digitizing standard ledger operations.
- Phase 2: Conversational Support (The Chatbot Era): The introduction of early chatbots allowed banks to offload basic inquiries, reducing the pressure on call centers.
- Phase 3: The Pilot Projects (The Proof of Concept): Credicorp began testing advanced conversational interfaces through Mibanco via WhatsApp, allowing users to initiate loan requests and open accounts through natural language.
- Phase 4: Agentic Integration (The Current Frontier): The bank is currently working on integrating these agents directly into the main BCP mobile application, moving beyond simple information retrieval into the realm of autonomous task execution.
- Phase 5: Hyper-Personalized Finance (The Future Goal): The long-term vision involves AI agents that act as proactive financial managers, offering advice on long-term savings goals—such as university funding—based on real-time data analysis.
Supporting Data and Real-World Applications
The efficacy of this model is already being validated in smaller, more agile segments of the Credicorp ecosystem.
Mibanco and WhatsApp
Mibanco has become the group’s testing ground for high-engagement AI. By leveraging the ubiquity of WhatsApp, the bank has enabled entrepreneurs to request loans, manage savings accounts, and receive business-related financial advice without ever leaving their messaging interface. The success of this model proves that when financial tools are integrated into a user’s existing habits rather than forcing them into a new, complex app environment, adoption rates and user satisfaction increase significantly.
The iO Experience
The iO digital bank, another component of the Credicorp ecosystem, has deployed a sophisticated conversational agent that allows users to perform longitudinal analysis on their own data. Users can query the AI to compare spending habits month-over-month, detect anomalies in their expenses, and receive personalized tips on savings. This serves as the blueprint for the more complex task-execution features being prepared for the flagship BCP app.
Official Perspectives: The Leadership Vision
The shift is being driven by a top-down mandate to make digital banking as intuitive as a human conversation.
Rafael Lemor on User-Centric Design
Rafael Lemor, manager of the Artificial Intelligence area at Credicorp, emphasizes that the bank’s strategy is a direct response to the pace of digital change. "We are constantly evolving our application because digital trends move at incredible speeds and client needs are in constant flux," Lemor stated. For Lemor, the goal is to remove the "friction of navigation." If a customer needs to perform a task, the technology should adapt to the customer, not force the customer to learn the software’s layout.
Patricia Wissar on Executable Interfaces
Patricia Wissar, Manager of Digital Channels and Transformation Capabilities at BCP, highlights that the "next stage" is defined by the verb to execute. "It is a conversation based on action, not just information," says Wissar. "The user experience will be defined by how the interface can accompany the user through the completion of a request." Wissar suggests that the interface itself is the new battlefield for competitive advantage in banking. If one bank allows you to "pay my tuition" via voice, and another requires you to click through five screens, the winner is clear.
The Implications: Safety, Security, and Autonomy
While the transition to agentic AI offers a frictionless future, it introduces significant challenges that the financial sector must navigate with extreme caution.
The Security Dilemma
Moving from "querying" to "executing" creates a new attack surface. If an AI can execute a wire transfer based on a voice command, how does the system verify that the voice is indeed the authorized account holder? How does it prevent "prompt injection" attacks where malicious actors trick the AI into moving funds?
Credicorp acknowledges that these capabilities will be gated by strict authorization protocols. The implementation of biometric authentication, multi-factor verification, and "human-in-the-loop" safeguards for high-value transactions will be essential components of the rollout. The bank intends to roll out these features with high-security guardrails, ensuring that while the interface is "simple," the security backend remains ironclad.
The "Black Box" Problem
There is also the matter of financial advice. If an AI agent recommends a specific savings plan for a user’s child’s university education, the bank is, in effect, providing automated financial advice. This necessitates a high degree of transparency in how the AI reaches its conclusions. Regulators are increasingly scrutinizing "black box" algorithms, and Credicorp’s development team must ensure that the AI’s logic is explainable, fair, and free from inherent biases that could lead to financial harm for the user.
Changing Customer Habits
The broader implication for society is the potential obsolescence of the "app." If an AI agent can bridge the gap between different services—perhaps connecting a banking app to a tax filing service or a budget management tool—the traditional concept of a "mobile banking app" may dissolve into a singular, conversational interface.
Conclusion: A New Era of Financial Literacy
The initiative by Credicorp and BCP is more than just a software update; it is an attempt to democratize sophisticated financial management. By removing the technical barrier of entry—the "navigation tax"—the bank is potentially opening the doors for millions of users to engage in complex financial planning that was previously reserved for those with the time or expertise to manage it.
As the development continues, the industry will be watching closely. If successful, the move toward agentic AI will prove that the future of banking isn’t in a faster, better-designed button, but in the total elimination of the button itself. For the consumer, this promises a world where the bank is no longer a place you go or a screen you open, but a service that lives in the flow of your daily life, always ready to execute, advise, and protect.
