When the GAFAs take over the banking relationship


« Hey Google, what’s my balance today? »

When the GAFAs take over the banking relationship

Imagine that we no longer use our computers or smartphones to access our banking information and perform banking tasks, and instead use the most natural channel available to mankind? Oral communication.

Communicate means, in its simplest definition, talking and trying to understand each others. Which is mainly achieved through language.

"The Major Transitions in Evolution" published in 1997 by John Maynard Smit a British evolutionary biologist and geneticist, and Eörs Szathmáry, a Hungarian evolutionary biologist makes reference to eight major transitions in the evolution of life. The most recent being language which is a human attribute only and participate in the creation of the foundations of our societies.

Why is this significant?

Voice recognition used as an input is one of the most efficient form of computing. That's what Mary Meeker, an American venture capitalist and former Wall Street securities analyst, thinks. In her 2016 Annual Internet Trends report, she already explained that people can express 150 words per minute when they can only type 40.[1] Meeker said “when speech recognition reaches 99% accuracy, people will go from barely using the tool to using it all the time[2].

At the moment, the champion in this field, Google, has reached the promising score of 95% in May 2017.

 What is the point for Banking services?

With the recent banking regulations that are shaking up all European banks (PSD2), new trends suggest the threat would not necessarily come from fintechs, however important they may be in the overall process of digitising and uberising payments and banking services.

Let’s remind PSD2 provides a common legal framework for the development of payment services for the Euro Area. In other words, the regulation aims at opening up payment markets to new financial players. We will focus on so called TPP (Third Party Provider) in that paper:

- The account information service providers (AISP):

Thanks to their licenses and customers authorizations, these players will have the ability to retrieve customers’ banking information and to display it to them in specific interfaces. This means for clients that they will be able to visualize all their accounts even if registered in different banks.

- The payment initiation service providers (PISP):

These actors will have the ability to initiate credit transfers on behalf of the client. Payers will initiate payments through the PISP which will carry the instruction to the bank rather than initiating payments directly to their banks.

 Well, let's get back to our subject

A year ago, we were mainly talking about threats related to Fintechs such as Bankin' or Lydia, offering already for users the functionalities to aggregate accounts and make payments.

Mainly because this new directive, PSD2, was initially created following repeated alerts from German banks about Sofort’s[3] activity.

As a result, little attention was paid to the GAFAs.

But, have you ever heard of Apple Pay or Amazon Pay?  By the end of 2018 in France, Google will make Google Pay available. And the latest one, still: Facebook Messenger Payments, which will be launched in the coming months.

 In PSD2 working groups, major banking players "confront" their points of views with those of the TPPs and one discussion caught our attention: the use of voice commands should not be neglected and should be made possible by banking APIs.[4]

These voice commands precisely, through which channels would they go?

Personally, we are only familiar with most known: Alexa (Amazon), Google Assistant, Siri (Apple)... And again: Facebook, which will use Alexa supplied by Amazon.

 Let’s imagine next steps:


 “Hey Google, what’s my balance today?” “Today, your balance is 4568,68€”

“Hey Siri, can I pay my rent of 800€ tomorrow at 9 am? “Sure. This is now saved as a future payment”

“Hey Alexa, can I do a 50€ transfer from my favourite account to my sister’s one?” “Done.”

“Hey Google, how much do I have on my shopping budget sub-account? “In your shopping budget sub-account you have 50,50€.”

This could be our interaction with our bank in the near future. But giving a closer look, these interactions would not really be with our bank. We will use a channel that has absolutely nothing to do with our bank: a personal assistant, equipped with voice recognition technology. We will no longer use banking applications, at least not directly.

 Is that really possible?

As per observed in OpinionWay  2015 study by SG, 90% of customer relationships are maintained through the Internet within banking services.

So let's take a closer look at the most common banking tasks, the ones we ultimately do online. According to a February 2018 study by SAB and CGI[5], online transactions realized by users are:

-        Account consultation (99% of respondents using a banking or web-banking application use this feature),

-        Transfer to another account (94%),

-        Budget analysis (78%) and

-        Subscription to a banking and financial product (45%).

 According to this same study, the exact same tasks can be carried out by voice recognition and assistance.  The consultation of account balances (51% of respondents are in favour), the request for information on banking products (55%), ordering, booking or purchasing (47%), the initiation of a bank transfer (35%) and authentication by voice to access their accounts (42%), are most quoted features by respondents panel.

The take away of these results? We assist to a growing maturity of digital behaviours: Online has gained a lot of ground, just as the Vocal is about to do.

 What could slow vocal technology growth ?

Along with innovation, and especially with voice recognition to manager accounts, perform payments… comes a real effort on communication and education in order to reach adoption.

One of the major obstacles is the lack of trust. Consumers are deeply mistrusting GAFAs regarding the use of personal data and this lack of trust is a core hindrance for mass adoption.

CGI and SAB findings show that, 3% of respondents seem very confident and 11% confident about GAFAs. (compared to 14% very confident for a bank and 44% confident). Nevertheless, we can notice that the level of trust in GAFAs is exactly the same as the trust placed in neo-banks. And we have all witnessed the rapid evolution in the use of neo-banks: Revolut (created in 2013) and N26 (created in 2015) have respectively reached two million and one million customers.

Another brake to market adoption is the security of data transfer and payment transactions. How to ensure the origin and the compliance of financial  transactions with as single input: a voice? How to authenticate a voice and make it a digital signature?

What PSD2 tells us about security and strong client authentication is clear:

Regulatory Technical Standards on strong customer authentication (Article n°4)

Where payment service providers apply strong customer authentication […] the authentication shall be based on two or more elements which are categorised as knowledge, possession and inherence and shall result in the generation of an authentication code.


So to strongly authenticate themselves to payment services, to initiate a payment or to access personal information protected from strong authentication, users can use a combination of Inherence and Possession to authenticate themselves to the bank.

Biometric technology is probably the most promising but also the most challenging method to implement. It is main advantage is that the user always has his "authentication code” with him because he is the password. No more risk for him to lose it or being hacked through password breaking.

 In short, we could thus imagine to couple smartphone platforms, voice recognition and/or finger prints to meet strong authentication required in PSD2 scope.

Last notable cause of tumultuous growth for voice recognition would be the Feasibility. Since 2017, the CNIL in France has authorized the implementation of devices for authenticating customers by voice recognition on an experimental basis. These devices are used to initiate online payment and to login into customers’ accounts. American and European banks are trying the implementation of voice recognition in partnership with the GAFAs: Ally Bank or Wells Fargo with Alexa in the USA and Banque Populaire, Caisse d’Epargne or Boursorama with Google Assistant in France, for instance.

Despite regulators’ timid steps forward, the GAFAs will continue to develop their own Artificial Intelligence to reach the 99% accuracy score in view of their research and development expenditures, according to the last 2018 analysis of Kleiner&Perkins[6].


What to take away?

Currently, a lot of research are influencing the implementation of new technologies that within banks but they are, for most, still in developing or testing phases.

In this use-case, banks can use PSD2 to provide greater user experience through a single application. The use of voice recognition through the GAFAs is a huge opportunity for banks to build strategic partnerships. Otherwise, GAFAs will end to become competitors, and serious ones (especially because we know the competitive advantage taken by the GAFAs regarding big data and personal data analysis). If personal assistants become the most common way to proceed banking transactions it will be tough for banks to maintain the customer relationships. Their prime position is at stake here.

Language and oral communication is what created our societies, according to our two biologists John Maynard Smit and Eörs Szathmáry. So, why can’t we see language and oral communication, for its simple and natural form, the best way to walk banking services further ?

About the author


Margaux Rudnicki, is consultant for Initio Luxembourg since June 2017. With more than 3 years of experience in banking consulting and 2 years in more marketing and digital strategy fields, Margaux was part of different type of project always related to change management, digital strategy, user experience, analytics, tools implementations and regulatory compliance linked to final user activities. As a Business Analyst and as a Project Manager Officer she worked specifically for projects in Retail banks and Private Banks, in France and in Luxembourg.


Marie-Charlotte Renaux. With 3 years of experience in Finance, Marie-Charlotte had the opportunity to get involved in transversal projects in Private Banking and Investment Banking. As a consultant she managed change within different types of organizations. 

She is supporting clients on complex challenges such as regulatory adaptations, project management, business operation models adaptation and actively participate to blockchain (DLT) researches and implementation

[1] https://www.kleinerperkins.com/perspectives/2016-internet-trends-report

[2] According to Five Highlights from Mary Meekers 2016 Internet Trends Report - Forbes

[3] An online direct payment service

[4] Summary of the 10th Meeting of the API Evaluation Group 25 June 2018

[5] « Services numériques dans la banque : Autonomie client & Digital Care Usages, opinions et attentes des Français » - February 2018 – CGI & SAB

[6] https://www.kleinerperkins.com/perspectives/internet-trends-report-2018