V2 recently built a customised experience and insight capability for an FSI client, enabling customers to effortlessly self-serve by leveraging AI behind the scenes, prioritising the right interface and information in real time.

The value to the business in building out dynamic AI capability versus a more serialised  Software Development Lifecycle (SDLC) pathway was clear. 

The build of the AI solution was a purposeful, well-defined, and lean approach that aligned with functional business requirements. Instead of ‘overloading the LLM with company wide data’ there was a need for AI to make independent decisions and respond to new situations.

The solution was integrated with internal systems with the ability to make external calls, both for optimal learning and responsive action. The system and its actions formed the foundation for agentic AI, acting autonomously, enhancing intelligence across business units, improving customer engagement and elevating the whole user experience. 

It is hypersonalised and actionable examples like this where agentic methods in AI are driving a new frontier in modernisation, disrupting traditional design and build methods. It is greatly increasing delivery velocity and also demonstrating how AI use cases with the right boundaries can augment across systems to form a new frontier in organisational intelligence.

What is Agentic AI and why is it so advantageous?

Agentic means AI ‘will act’ and not just ‘assist’. Agentic workflows run themselves, making decisions and completing certain tasks without waiting for input.

Agentic AI adapts dynamically, is self-learning and is capable of proactively optimising tasks based on real-time data.

Agentic AI will become the game changer in Banking

If you don’t build Agentic AI into your business, you can be sure you will soon be competing against a business that has.  

AI is becoming core to modernisation and particularly visible in the Retail Banking sector, where Payments, Cards, Products, Customers and Transactions are being disrupted.

There are numerous examples of AI currently being utilised by leading banks today - chatbots, fraud detection and prevention, customer 360 platforms, reporting and insights. These capabilities help remove bottlenecks and heavy lifting of tasks, dramatically reducing queues and wait times.

However, while AI tools and capabilities can offer large improvements, they are currently bringing incremental changes that are affecting return on investment (ROI). They may fall short of delivering real value if they cannot dynamically optimise those processes. In contrast, by making decisions using logical reasoning and keeping sound operations, Agentic AI with agent deployment streamlines operations where benefits can be truly recognised at scale.

We are seeing a broad set of use cases developing within Retail Banking

These range from hyper-personalised experiences, autonomous financial planning, real-time fraud detection, and intelligent automation of back-office operations.

In Transactions - The implementation of agentic AI systems focuses on monitoring transaction processing. The system learns typical account activity patterns to detect meaningful deviations, which could signal errors or fraud, then flags these for review while decreasing manual intervention times and supporting regulatory compliance.

Across Payments - AI-driven automation is streamlining real-time payment processing, fraud detection, and risk assessment, ensuring more secure transactions for customers.

Using Cards - Agentic AI is coming into play with card management by analysing transaction attributes across spending velocity, location shifts and merchants. Automating responses like blocking cards, notifying customers and generating alerts for credit limits or initiating card reissues.

With Customers - AI-driven customer profiling enables hyper-personalised banking experiences by adapting user profiles dynamically while linking them to customised product recommendations and marketing campaigns that draw from transaction history, app usage patterns, declared financial objectives and financial observations. 

 Finally, Products and Agentic AI will converge shortly to dynamically adjust financial product offerings based on market trends and individual customer needs, ensuring banks remain competitive and relevant in an evolving landscape.

Despite the initial implementations and further potential, there are challenges to be met.

The key challenges in Agentic AI mimic those of typical AI, namely: Privacy, Security, Cost and Integration. All implementations of AI must align with compliance and regulatory requirements, which can be challenging, whilst also creating a level of trust and transparency with customers.

Agentic AI with further optimisation and action in business processes, requires careful oversight of AI models and resulting actions - with monitoring, alerting and testing at the forefront of business risk management.

Data continues to be paramount, driving real time insight and the next best action. Ideally, data has an enabled, federated environment with strong data governance and quality. 

The integration with legacy systems may need a level of abstraction to interoperate with AI, whilst also transforming that underlying systems layer. 

What's Next for Agentic AI in Retail Banking

Agentic AI is a large focus for the team at V2, as we believe it is the next frontier in banking and innovation for our customers. It is set to evolve in transformative ways in a very short period of time.

We believe that AI will feed a virtuous cycle. As the lines blur between traditional retail banking and FinTechs, AI will take on a larger role in the ecosystem. Traditional retail banking services will be embedded into non-financial platforms like retail sites, apps, and even social media platforms, offering financial services like loans, payments, and savings accounts. Agentic AI will further underpin and accelerate compliance, regulation, efficiency and customer experience.