Tuesday, March 20, 2018

Trends & Challenges : Conversational Banking


Three Trends Impacting Growth of Conversational Banking

According to Accenture, three trends are coming together to support the transition from transactions to interactions.
  1. Messaging is now the preferred method of interaction: WhatsApp, Facebook Messenger, WeChat, Telegram, Snapchat, etc. have overtaken social media as the preferred way of communication on mobile devices. They are both simple and intuitive, leveraging text or voice-based interfaces. These apps also are AI-ready, offering easy integration with chatbots and cognitive agents.
  2. AI is now accessible to all organizations: The decreased cost of data storage, analytic tools and development in machine learning and deep learning enables the automation of repetitive customer support tasks, and lower level advisory services. Over time, AI will enable more advanced interactions at even lower costs.
  3. Mass personalization is now possible: The intersection of big data, advanced analytics, and predictive models is enabling personalization at a mass level. This supports vastly improved consumer experiences while increasing expectations of all financial services providers based on the delivery of services across industries.

Challenges to Transitioning to Conversational Banking

There are already more than 33,000 bots on Facebook Messenger and more than 100,000 bots offered by Chinese digital giant WeChat. These bots offer customer support, e-commerce guidance and other interactive experiences.


While a good first step, most bots today have limited capabilities. Some simply replicate already outdated FAQ experience, while others simply replace low level human-based customer service capabilities. The ultimate goal is to move more interactions to voice, supporting sales, advanced customer care and advice, with greater personalization, a better UI, less friction and more automation.

This is a unique example where the technology and consumer acceptance is already out there, but strong use cases have yet to come. Many voice capabilities have had initial success, only to have growth stalled due to lack of value-added capabilities or consumer education. Organizations will need to use education and value-added benefits to ‘sell’ them on new ways of interaction. Without education and consumer-centric benefits, it’s unlikely consumers will use a chatbot, let alone digital assistants as opposed to current human channels.

Another barrier to transition to conversational engagement is the siloed nature of banking. Conversations interactions will not move forward if the experience is not seamless, with all insights, from all areas of the organizations arranged around the consumer. Traditional segmentation or product ownership silos will only inhibit natural and successful dialogue.

Conversational Banking Strategy


Six Keys to a Successful Conversational Banking Strategy

When building a strategy for conversational banking there are six components Accenture believes are necessary for success. Many will require a complete restructuring of current back-office processes, whil others are less daunting.
  1. Acquire Needed Talent: Advanced capabilities and skills including neuro-linguists, voice recognition experts, AI developers, and CUI designers are some of those most in demand today. Talent will continue to be a barrier for many years as demand is far outstripping supply.
  2. Understand Technologies: Understanding and acquiring new technologies will require a move to real-time back-office processing and ongoing monitoring of advances in text and voice-based technology.
  3. Improved Privacy and Security: The collection and application of expanded consumer insights requires a greater focus on securing this insight. Voice biometrics will also need to be utilized for conversational interfaces.
  4. Transparency: As chatbots and conversation interfaces better replicate human interactions, it will become more important to let consumers know they are interacting with new technology … and not a human.
  5. Consistency Across Channels: It will be important to allow for consistent experiences across channels. Selection of what channels to support may revolve around security as much as anything.
  6. Consider a Hybrid Model: Since AI-based interactions are not perfect and a great deal of testing will still be required with conversational interfaces, a hybrid model including technology and humans is recommended.



Banks Are Disrupting Payments – Two International Models

1) Open model - where banks build and control the processing infrastructure, but leave the front-end competition to the market.


2) Where banks are uniting into consortiums to control both parts of the business, processing and user interface, through their own solutions.


Modular framework (central banks are offering robust infrastructure)

Among the most impressive machines falling into the first model is China’s Internet Banking Payment System (IBPS), a part of an established second-generation payment system developed by the People’s Bank of China.


“IBPS connects the online banking systems of the various commercial banks. Payments between banks can be received in near real time, with acknowledgments within 20 seconds.


China offers credit transfers (push payments) and considers direct debits are considered as pull payments. IBPS is ISO 20002-based and is the latest service to operate on the foundation of the China National Advanced Payment System (CNAPS II) interbank clearing platform.


Over recent years, the use of IBPS has skyrocketed. By the end of 2016, there were 195 institutions with direct access to the online payment interbank clearing system. In 2016, the internet interbank settlement systems handled around 4.453 billion payments to the tune of about 37.46 trillion yuan. This represents a growth of over 50% in volume and 35% in value.”


FIS reports that going forward, the expectations are that banks and more interestingly third-party organizations will develop new and innovative products & services on top of the IBPS platform, leading to higher market penetration and dominance.




China’s IBPS is reminiscent of the NPCI UPI model, where the state builds a universal infrastructure, seamlessly powering payments across both sides of the industry – banks and compliant third-parties.


UPI is a platform designed for the mobile age that helps with easy integration of various payment platforms. UPI is powered by a single payment API and a set of supporting APIs. It has a fantastic value proposition including a simple authentication process, simple issuance & acquiring infrastructure, national interoperability, and more.


UPI has reshaped entity-entity relationships (whatever entities those are – merchants, payment gateways, banks, network owners).


As Harshil Mathur, Co-founder & CEO of Razorpay, explained, “…instead of having to build 57 bank relationships a gateway has to build right now, it needs to build only one relationship with the UPI. So it will ease our efforts and help out the merchant.”


Australia’s NPP is also a very important example of how institutions are redefining payments framework on a national level.


As described by Nathan Lynch, Regional Bureau Chief, APAC, Financial Crime & Risk – Thomson Reuters, the Reserve Bank of Australia had insisted that the NPP platform be designed using a distributed layered architecture. This meant the system would be divided into a Basic Infrastructure (BI) layer and have a range of Overlay Services (OS) that could sit on top, allowing innovation and flexibility at the user interface level.
“The BI consists of distributed connectivity points (banks), message flows, a switch, a fast settlement service, and an addressing service. The OS, meanwhile, promises to allow NPP participants, approved third parties and “FinTechs” to develop payment services that sit on top of the basic layer. This forward-looking design means the platform can evolve while the underlying architecture remains stable.”

Infrastructure + front-end competition with proprietary solutions

The second model is a bit more radical than the first one: bank-owned front-end as opposed to the back-end infrastructure (or, in other words, the modular model).


There is one well-known – Zelle in the US – and another one that just made its debut in the news – Japan’s MoneyTap.


This Monday, Ripple announced that the Japan Bank Consortium (comprised of 61 banks covering more than 80% of all banking assets in Japan) will release a smartphone application called “MoneyTap” – powered by Ripple’s blockchain technology – to allow customers of the bank consortium to settle transactions instantly, 24/7. MoneyTap is the first mobile app of its kind to be developed and used by multiple banks in the country.


Three members of the Japan bank consortium: SBI Net Sumishin Bank, Suruga Bank, and Resona Bank will be the first to go live on the mobile app in autumn of 2018. This will be followed by a staggered roll out to the rest of consortium.


MoneyTap allows the bank consortium customers to make instant domestic payments and only requires a bank account, phone number, or QR code. What’s more, Ripple shares, MoneyTap helps shed the costs associated with existing banking and ATM fees that are currently applied to domestic money transfers in Japan, making those payments not just faster, but cheaper overall.
Zelle is the US equivalent of what Japan is doing.


In 2011, Bank of America, Wells Fargo, and JPMorgan Chase teamed up to work on a digital payments solution that would allow their customers to send money to each other.


The first product they developed, with the help of bank-owned company Early Warning, was called clearXchange. The solution was officially rebranded as Zelle in June 2017, with an emphasis on bringing as many financial institutions onboard as possible.


Under the covers, Zelle still functions by combining a directory of emails and phone numbers matched to bank account data along network rules for moving money along the ACH network. Thanks to bank agreements, the money is made available immediately to the receiving user, despite the overnight timing of the actual money movement.


With eventual cooperation from more than 30 of the leading financial institutions in the United States and the world, Zelle is successfully gaining ground. About $1.5 billion worth of payments were processed through Zelle just in October 2017, which is up 90% from the same month last year. At that time, Zelle counted 2.5 million active users, with thousands more signing on every day. At the end of Q3 2017, BofA had 23.6 million active users of its mobile app. The volume of mobile check deposits made on the app corresponds to the work of 1,100 branches.


There are significantly many more examples we could cover – variations of those models and their mix exist across a large number of jurisdictions. But all of this to make an important observation – banks are taking full control over payments. Here is why it’s important: the reality is that financial institutions in developed economies service a large part of the national population, and in economies with a significant number of underbanked, they are looking to offer the easiest way to get plugged into the world of digital financial services. Both models have their advantages and challenges for different types of economies, but one thing is clear – they are run or majorly controlled by banks, not by startups or processors.