Customer case
Seeking productivity gains to increase the flow handled by conversational AI
About this client.
The client we're working with is the Corporate Investment Banking (CIB) arm of a bank with an international client base. This division serves 18,000 corporate clients and has 36,000 employees across 55 countries. To decongest its customer service on one of its platforms, the division decided to leverage the conversational AI tool already available within the Group.
HOW DID YOU APPROACH THE PROJECT?
Positioned at the interface between the business and IT teams, Talisker contributed to progressively and incrementally delivering more value with each new version of the conversational AI by designing a customer experience aligned with the needs of target users. To achieve this, three successive versions of the conversational AI were deployed:
V2, offering a first level of self-care to users by leveraging available user documentation (videos, one-pagers, user guides, etc.).
The AI was opened to the first pilot users (~300 users).
V3, enabling the AI to autonomously answer certain questions and assist users in ticket creation. For example, it could automatically retrieve all information related to a failed payment, allowing for a precise ticket to be created without any effort from the user.
The number of pilot users was extended to 600.
V4, marking a significant product maturity shift, with over 40 recognized intents and more natural conversations.
This version came with the start of a mass roll-out for EMEA clients, targeting 200,000+ conversations per year in the long term.
SETUP
1 Talisker consultant
WHERE DID YOU START FROM?
After a successful first mission with Talisker that led to the implementation of the conversational AI Amelia, the Corporate Investment Banking division—convinced of the benefits of conversational AI in alleviating daily operational tasks—decided to take things a step further by aiming to handle part of the client requests end-to-end through the conversational AI.
In this context, Talisker was once again engaged to structure the value creation for the end users of the platform.
The first version of the tool implemented—still at the Proof of Concept (POC) stage and not yet available to clients—was designed to redirect tickets to the appropriate service desks, based on the topic of the request and the user’s country.
IN YOUR OPINION, WHY WERE YOU SUCCESSFUL?
Talisker’s shared experience with the Amelia teams helped streamline communication as much as possible and ensure the right expertise was brought in at the right time, accelerating the project and empowering the teams involved.
The strong involvement of the business Product Owner for the conversational AI within the Corporate Investment Banking division, along with their significant drive to engage sponsors, was undoubtedly a key factor in the success of the mission.
The thorough use of rich, pre-existing client-facing documentation also played a crucial role in promoting self-care, by enabling the rapid delivery of high-quality content to clients.
AND WHAT WAS THE SITUATION WHEN YOU FINISHED?
The structuring of continuous improvement and regular monitoring of user conversations on Amelia made it possible to identify a wider range of request types. As a result, the AI’s recognition scope expanded, allowing it to handle more intents and thus deliver even greater value.
Organizational work led to the creation of a product team composed of three business resources and four AI-dedicated resources for the Corporate Investment Banking division.