Speaker: Rohit Dhawan, Executive Director and Group Head of AI, CDAO - Group COO, Lloyds Banking Group

Astitva Karinesh from Lloyd's Banking Group elaborated on the concept of agentic AI by differentiating it from general AI, explaining how it moves from basic foundation models to more sophisticated orchestrating agents that integrate multiple specialised tools and models. This interconnected system allows for dynamic reasoning and real-time actions, enabling more accurate and complex responses to queries, which traditional deterministic routes cannot handle. The idea was illustrated using examples from the banking sector, showcasing how agentic AI could revolutionise customer interactions and service delivery, leading to more personalised and efficient experiences.

Katie Adams continued the discussion by emphasising the practical implications of agentic AI for the future of banking. She outlined Lloyd's Banking Group's approach to integrating AI into their operations, focusing on generating immediate business value, enhancing developer experiences, and democratising AI knowledge across the organisation. Adams provided a fictional scenario involving a customer named Sarah, who uses an AI-powered financial advisor to navigate the home-buying process. This example demonstrated how agentic AI can simplify complex tasks, such as obtaining a mortgage in principle, by leveraging interconnected models and real-time data. The session underscored the importance of reimagining business processes rather than merely optimising existing ones, advocating for a fundamental shift in how businesses operate.

The session concluded with a discussion on the key considerations for implementing agentic AI in business. Adams highlighted the need to carefully evaluate when AI is the appropriate solution for a problem, as not every issue benefits from AI intervention due to its complexity and unpredictability. She stressed the importance of reimagining processes with AI rather than just automating inefficient ones. Additionally, the necessity of establishing robust control mechanisms and guardrails to ensure AI systems act in the best interests of customers and comply with regulatory requirements was emphasised. The speakers conveyed a sense of optimism about the transformative potential of agentic AI, envisioning a future where AI significantly enhances customer experiences and business operations.

Takeaways

Agentic AI represents a significant evolution from traditional AI models

Unlike general AI, agentic AI integrates multiple specialised tools and models, allowing for dynamic reasoning and real-time actions. This interconnected system can handle complex queries and provide more accurate and personalised responses, which traditional deterministic approaches cannot achieve. The concept was illustrated using scenarios from the banking sector, showcasing its transformative potential for customer interactions.

Implementing AI requires reimagining business processes, not just optimising existing ones

Katie Adams emphasised that applying AI should go beyond automating current workflows. Instead, businesses should fundamentally rethink how they operate to leverage AI's capabilities fully. This approach can lead to more innovative and efficient services, creating significant value for both customers and businesses.

Robust control mechanisms are essential for successful AI implementation

Adams highlighted the importance of establishing strong guardrails and monitoring systems to ensure AI acts in customers' best interests and complies with regulatory requirements. These controls are crucial for maintaining trust and ensuring the information provided by AI systems is accurate, useful, and ethically sound.