By Chris Kapfer
While steering the calls, Eva reduced incorrect routing by 73% and has also resulted in overall 50% reduction in the time taken to digitally serve a customer as compared to IVR.
Emirates NBD Virtual Assistant (Eva) is the first-in-region natural voice recognition technology based automated phone banking assistant and also first-in-the world for Arabic language. Eva freed customers from the complex and frustrating maze of the touch-tone IVR; saving them considerable time and effort to get a solution for their issues.
IVR is increasingly reaching its optimisation levels in regards to digitisation and scalability, however the technology has never been well accepted among banks. It has huge limitations in the number of automations it can offer and the number of services that can be added to it. At the same time, it faces challenges with high call handling time, high abandoned calls and low customer experience, hence driving up operational cost and leading to brand and value erosion.
The bank’s key operational reason to choose phone banking optimisation as a starting point was to arrest increasing calls to agent due to a reduced digital containment rate, an indicator that measures the degree of end to end digital self-servicing.
EVA was designed in the lines of personal assistant like Apple's Siri, Google Now and FaceBook's M, however it has been enabled with full back end integrations for end to end banking fulfilment capabilities.
Eva is trained to deal with 150 skills which are the most common reasons why a customer calls in and this can be scaled further as needed. Conceptualised in 2015, Eva was launched in February 2017 and implemented across all major retail financial services helplines by June 2017. The assistant currently handles around 500,000 customer calls per month with a Recognition accuracy of around 90% for English and 86% for Arabic.
Based on technology provided by Nuance, with EVA, IVR is not eliminated as it continues to run in the background, but the customer navigates through the channel by using simple voice commands without the need to press buttons and hear long winding menu mazes, a practice still common across domestic banks in the Asia Pacific and the Middle East. With EVA, the IVR tree does not exist in the customer interface anymore.
In addition, there is a seamless machine to agent handshake as the calls are routed, hence the agent is aware of the intent of the call before he addresses the customers. Eva increased automation and improved end to end fulfilment on phone-banking from 11% to 19% while reducing the calls to agents by 14%. While steering the calls, Eva reduced incorrect routing by 73% and has also resulted in overall 50% reduction in the time taken to digitally serve a customer as compared to IVR.
Before launching EVA, the bank had about 20 automation services. The current platform has 47 fully automated services. The most significant improvements were witnessed by the reduction in the number of abandoned calls by 11%. Positioning for the right agent to the customers is a key challenge for any bank but with EVA, the bank could reduce wrong agent skill routing and thus recalling of the customer by 73%.
The assistant also generates digital leads for cards, loans and account products; and the bank is in the process to add the capability of fulfilling the immediate conversion of a 0% instalment plan for credit cards in near future.
Source: Emirates NBD
Banks are tuning and adapting their machine learning capabilities which will take time to learn. The key business case is often not efficiency but scalability and the fact that financial services providers have customer services solutions at hand that can be rolled out quickly.
SEB, Sweden (Aida, IPsoft)
SEB, the leading Nordic bank launched Aida in 2017 with a chatbot based interface that handles external service requests for more than a million customers.The bank’ use of a ‘cognitive agent’ began as a pilot project for company’s internal help desk in September 2016. Aida handles customer queries like password resets, step-by-step assistance with credit and debit cards, banking location services, and ID verification. It has reached a 90% accuracy rate in understanding and completing tasks while the bank set a 30% referral target to human agents who can focus on more complex interactions.
GarantiBank, Turkey (MIA, Nuance)
The bank’s mobile interactive assistant (MIA) launched end of 2014, the iGaranti virtual assistant leverages on natural, conversational language to inquire about account details, transfer funds , get exchange rates, buy/sell foreign currency, cancel cards via seamless machine agent handshake. Garanti reports that approximately 60% of its app customers are using MIA. Based on the same platform it developed an ‘Empathy Assistant’ to guide branch employees which provides automatic diagnosis of the root cause for the customer issues and supports the employees with a list of actions to resolve them.
Source: Asian Banker Research
Emirates NBD is one of first banks in the Middle East to launch voice based virtual banking assistant and first in the world for Arabic speaking voice assistant. The initiative has resulted in significant improvement in customer service as evident from reduction in call abandonment, first call resolution and IVR containment. More importantly this is a scalable solution that allows the bank to introduce new service features at its call centre which it could not do previously because of limited menu capabilities of IVR. Yet, this is an assisted learning technology that will need time and resources to be constantly developed to reach it potential.