Designing conversational AI strategy for a big international bank

Overview
We partnered with a global financial institution to create their modular approach for conversational AI search. The goal was to design a seamless and personalised experience for customers, which would increase their autonomy and reduce support costs.
The Challenge
The bank had experience with conversational AI, having launched and iterated upon their customer and employee-facing chatbots. However, the release of advanced multi-modal models like GPT-4 presented an opportunity to progress beyond scripted interactions and build a more intuitive and natural experience.
They sought to create a strategy that unifies the experience of Search and Help with the support of conversational AI. Additionally, they aimed to leverage their newly adopted modular design approach.
Our Approach
Our team worked closely with the bank’s internal conversational design team to create a modular approach for conversational search. Here's how we achieved this:
  • Set the foundation: we began by evaluating the existing customer experience, identifying available channels and user options. Simultaneously, we explored how leading competitors in the banking sector offered search and help functionalities to understand customer expectations.
  • Understand Expectations and Goals: we facilitated a workshop with the project’s main stakeholders to agree on goals, desired outcomes, expectations, concerns, assumptions, and "nice-to-haves."
  • Detail project approach: we created a detailed timeline outlining key tasks, milestones, and dependencies. This ensured stakeholder expectations were managed.
  • Design and test conversational modules: we generated a range of design concepts based on insights gathered from market analysis, process mapping, internal data collection, identification of top use cases, and prototyping. User testing, conducted online and in-person with bank customers, played a critical role in validating our assumptions.
  • Assess feasibility: we held parallel sessions to assess the feasibility and technical implications of the proposed solution, ensuring the best possible client experience.
  • Recommendations and Next Steps: the resulting strategy aimed to incorporate the insights we generated into an evolutionary approach - gradually building upon our partner’s existing capabilities towards more advanced conversational experiences. The implementation of these recommendations is ongoing.
Our Solution
We co-created a modular solution and strategy that integrate search and conversational functionalities into a single, seamless user experience. The solution envisions an omni-channel, omni-present central entry point for customers to search, ask questions, filter, navigate, or solve problems around cognitive, complex and onboarding tasks related to online banking.
Key features of our solution:
  • Speech-to-Text: enables the convenient capturing of customer intent in a natural way.
  • Fast & Accurate Intent Recognition: provides reliable service and allows customers to use use their own words.
  • Entity Recognition & Context Analysis: understands the user's input and determines the appropriate API to connect with.
  • Response Generation: delivers the customer an answer, suggestion, or quick actions.
We designed two additional modules to achieve this:
  • Suggestions: offering a range of different options based on the customers input as they type. After submitting, the suggestions module can offer alternative options if the primary answer is not sufficient.
  • Dialogue: combining different systems in order to give more specific, tailored and complete answers to the customer.
Benefits

Increased Speed
‍

delivers relevant information faster, reducing navigation back and forth.

Enhanced Customer-Centricity

allows users to express themselves naturally, without being restricted to banking terminology.

Improved Flexibility & Scalability

the solution's non-linear flow fosters flexibility, scalability, and better adaptation to specific customer segment needs.

Increased Engagement
‍

personalised assistance encourages users to explore additional features and services.

Reduced Support Costs
‍

customer reliance on human assistance decreases, as most of the queries can be handled through the conversational search.

Customer Insights and Data

can be improved with up to 60%* as these systems can analyse user inputs more comprehensively, providing deeper insights for personalisation and service enhancement.
conclusion
Together with our partner, we successfully created their modular approach for conversational AI search. The approach paves the way for a future where customers can enjoy a seamless and personalised experiences. As such, they can manage their finances more autonomously while reducing the burden on support teams. Our approach also allows for flexibility and scalability, ensuring the platform can adapt and grow alongside the bank's needs and customer expectations.
FESTINA LENTE

Curious to learn more?

We would love to explore how we can partner with you on Data & AI initiatives for the benefit of your organisation, people and society

Reach out to us!
Designing futures for all of us.
AboutServicesWorkInitiativeInsights
human@deus.ai
BlogResearchPodcastJoin usContact us
DEUS Amsterdam, Lauriergracht 112-C 1016 RR Amsterdam, NL
‍
DEUS Porto, Rua Roberto Ivens, 288, 4450-247 Matosinhos, PT
‍
DEUS A Coruña, Calle Ciudad de Lugo 35 - Bajo A 15004 A Coruña, ES
Subscribe to our newsletter and we’ll keep you posted on our research, podcast and other AI goodies.
Thanks for subscribing!
Something went wrong with the submission.
© 2024 DEUS.AI. All rights reserved.
Privacy Policy
Our site uses cookies to enhance the viewing experience. By continuing to browse our site, you agree to our Privacy Policy.
Ok, I got it
Designing futures for all of us.
Subscribe to our newsletter and we’ll keep you posted on our research, podcast and other AI goodies.
Thanks for subscribing!
Something went wrong with the submission.
AboutServicesWorkInitiativeInsights
BlogResearchPodcastJoin usContact us
DEUS Amsterdam, Lauriergracht 112-C 1016 RR Amsterdam, NL
DEUS Porto, Rua Roberto Ivens, 288, 4450-247 Matosinhos, PT
DEUS A Coruña, Calle Ciudad de Lugo 35 - Bajo A 15004 A Coruña, ES
© 2024 DEUS.AI. All rights reserved.
Privacy PolicySustainability