Coming into 2024, a key theme is emerging: banks are progressing from just experimenting with generative AI and towards spinning up production projects.
A staggering 70% of CIOs in the banking sector view generative AI as a ‘game-changing technology’.1 This enthusiasm is reflected in the adoption plans, with 55% of CEOs aiming to deploy generative AI within the next two years and 25% of banks reporting that they have implemented and are scaling their generative AI initiatives.2
But there’s often a missing piece of the puzzle required for those applications that can be quite difficult to solve: how do you ensure these experiences are personalised and powered by a true understanding of each individual customer?
The RAG pattern in GenAI Apps
One of the emerging patterns in this landscape is the use of retrieval-augmented generation (RAG). This pattern isn't just about leveraging large language models (LLMs); it's about enriching these models with precise, contextual data that pertains directly to the query at hand.
Bud: providing customer context
Transactional data is the richest source of customer data that banks possess. But this data is hard to get the full value from due to its difficulty being interpreted and classified in ways that make it usable for your business. As such, it’s rarely used in automated systems outside of fraud and AML, or by stretched data analysis teams for specific purposes.
At Bud we use LLMs that we have specifically architected, trained and refined for over six years. We make transactional data clean and easy to use for complex and regulated use cases that span areas such as credit risk, personalisation/marketing, and fraud/AML. We then use that enriched, explainable, and human-friendly transaction data, as well as other bank data sources, to build a comprehensive 360-degree view of a customer.
This gives both a summary of current characteristics and actionable insights, helping banks not just understand but also predict customer needs and we can deliver that as context straight into your Gen AI Applications today. By combining this with transactional data, you can ensure results are smart but also that such conclusions and suggestions are both explainable and predictable.
Customer context in your GenAI applications
Whether you’re building internal tools for your marketing departments to manage campaigns or chatbots and virtual assists for in-app experience or customer agent efficiency – customer context can’t be considered an optional extra.
66% of financial executives believe that hyper-personalisation, fuelled by advanced data analytics, is the future of competitive differentiation in banking. Over 75% of financial leaders say their organisation is likely to leverage generative AI in empowering more sophisticated chatbots and virtual assistants, or predicting it will be used for both customer service and personalisation.3
It also cannot be left to a second phase. A significant 61% of customers desire personalised recommendations,4 and a staggering 62% would switch financial institutions if they felt they were being treated impersonally.5
Early adopters with Bud are already reaping the benefits of this, and not just in metrics such as NPS and churn.
One retail banking client of Bud’s saw a dramatic increase in product upsell success, from 49% using traditional methods to 76% with Bud's AI platform while another global bank witnessed a 345% surge in total feature visits after implementing Bud's AI-powered experiences and a 10% reduction in call centre contacts.*
Personalisation is no longer a luxury, but a necessity in the banking sector. In your GenAI apps, that means utilising strong customer context.
The platform approach
Currently, we see building customer context for GenAI and other applications being a hot topic across multiple departments and teams at banks with applications being limited to narrowly defined processes. This allows for a quick start, but soon a need for a more orchestrated customer context infrastructure approach will become evident.
It is not only a matter of dealing with the complexity of providing proper customer understanding across an organisation and the inefficiencies of multiple teams tackling the same problem, but also that it is critical to consider a consistent view across the whole organisation.
For raw transactions and account balances this is guaranteed by the core banking platforms and a similar approach for the customer context will likely be required. Only this will guarantee consistent user experience and decision making across the whole organisation.
Contextualise your GenAI applications with Bud
As banks prepare to deploy or expand their GenAI capabilities, incorporating Bud’s data enrichment solutions can play a pivotal role in their strategy. By prioritising context and relevance, Bud helps its clients not only meet but exceed the evolving expectations of their customers, paving the way for a smarter, more personalised banking experience.
Interested in staying up-to-date with all things AI and Bud?
Sign up to our monthly newsletter!
* From anonymised customer stats and testimonials
1 eBankIT (2024), Uncharted Territory: Digital Banking Trends and Predictions
2 PYMENTS (2024), Banking on AI: Financial Services Sector Harnesses Generative AI for Security and Service 2024
3 KPMG (2023), The Generative AI Advantage in Financial Services
4 UK Finance (2023) The Impact of AI in Financial Services
5 ebankIT (2024) Uncharted Territory: Digital Banking Trends and Predictions 2024
