Voice assistants in financial services are getting ready

There is no shame in confessing that you have wished your voice assistance could answer some of the pressing questions that trouble we humans. 

When will I find love? When will I die? Or, maybe, Will I get the job? If you have a voice assistant, you have insulted it too, we bet. And there is no shame in confessing so either. Also, because the voice assistant won’t judge you. You have already realised that, at least subconsciously. 

But, have you yet asked her or him, whichever you prefer, to transfer money from your bank account? 

Under the umbrella of voice technology, there is one that stands out, the virtual voice assistant. By combining speech recognition—that provides the means for our speakers to hear our voices, and natural language processing—an AI subset that enables language understanding and speech synthesis, voice assistants help us to find information, play our favourite song, know the weather forecast or make an online purchase. 

We need to also price in another factor that is boosting fast developments in the voice technology field, computing power, because of its ability to process massive amounts of speech data very fast. 

Speech recognition is on the verge of emulating humans’ capabilities, as other AI-powered technologies are. This situation raises, very often, this question (sometimes untold but still hovering in the air): are we ready for this?  

Sherry Turkle, a prominent professor at the Massachusetts Institute of Technology (MIT) suggests that “we’re in the robotic moment”, namely, the moment when we are prepared to make “the robots”—please, allow us for that charming and simplistic generalisation—part of our lives.

The case of voice assistants is a clear example. It’s predicted that more than one in two households—55% to be precise— will own smart speaker devices in 2022, and that’s only a few months from now.. But are we ready to grant them access to our finances too? 

Read on. This article is the summary of an on-the-phone conversation we had with Andreas Braun about where we stand in the adoption of voice assistants in financial services and what the future for them looks like. 

Good to know!

Almost 60 years ago, in 1952, Bell Laboratories designed the Audrey system, the first speech recognition system that the world witnessed being born.
Only focused on numbers, it was able to understand a single voice speaking digits aloud. 10 years later, “Shoebox”, an IBM technology, could understand English words, 16 to be precise.
Since then, cumulative innovation-driven developments happened decade after decade but it was in the second half of 2000s when the technology skyrocketed. Thanks to the massive adoption of smartphones and, therefore, the use of speech recognition apps developed by GAFA (the “tech giants”), millions of people have helped the technology to take leap-frog steps and get perfected.

 
What organisations need to do to embrace voice technology

Speed, cost-efficiency, the feeling of not being judged, the ability of the robots to read our moods, multitasking performance, those are some of the advantages of using voice assistants. When thinking of a tangible example in the financial services industry, for instance a bank, voice technology embedded in different applications or devices offers additional advantages. 

Apart from the saving factors—cost and time—with banks and clients benefitting alike, other advantages such as availability (no need to stick to opening hours), convenience for clients (no need to dress up, make up or shave, for instance) and, most importantly, accessibility—that brings the possibility of attracting a broader audience, make voice technology seductive. 

A voice assistant might become the main choice for financial interaction for certain groups. People who prefer virtual interactions with machines rather than with humans, individuals with reduced mobility or limited visual capabilities, expatriates whose local language command is narrow, workers with a tough time schedule or working in the evenings, to name some, could access certain financial services quickly and conveniently. 

To capitalise on the technology more successfully, financial services organisations want to think of: 

  • The client experience should always be first. Strive for creating as effortless experiences as possible that bring true value. What do you want clients to feel when using any voice technology you implement? 
  • The strengths and best use cases for voice technology, i.e. assess which voice assistants and enabling devices are being used, how they are used and what the clients’ engagement with them is.
  • The accelerators or opportunities for voice to have massive growth, for instance, what the demographic characteristics of a financial institution’s clients are.
  • Clients’ concerns. Research what clients need to trust this technology and use it with confidence. Almost unavoidably, this type of solution needs to be brought in by means of a well-thought onboarding strategy. This strategy should also tackle the fear of losing privacy. We’ll focus on these matters a bit later.
  • Your data readiness. To make the most of voice technologies, you need to develop and maintain a solid data infrastructure and a documented, formalised data strategy. Equally, your business approach to AI adoption and where you stand is key. Otherwise, supporting voice-based solutions will be nearly impossible. 

However, before the considerations mentioned above, there is one first step that will uphold any voice-related venture: deciding the role it will play in the entire client or customer life cycle, where it is necessary and where it is not. 

Because, aren’t you considering voice assistants as an alternative to going to a business branch? We come back to this idea some lines below. 

 
We’re indulgent towards voice assistants

According to Andreas, the reasons why people like voice assistants are manifold. Obviously, he says, there is the initial interest and curiosity that has been driving a significant uptake of voice technologies in the last five years. 

There is convenience as well, a factor that ranks very high. 

It is nice and comfortable to manage your home from your couch without using remote controls. Or your car without having to focus on your touch screen which also increases security. 

On the other side of the coin, the technology is not quite there yet, considers Andreas.To him, it’s hard to argue that immediacy is a strong reason to use voice assistants. And in terms of efficiency, i.e. how many functions we can trigger when using an interface in a certain amount of time, voice isn’t yet (or, maybe, it will never be) a strong competitor to more “traditional” digital interfaces. Also, typed searches provide more functionality and fine-grained control to the user and the exclusion of certain terms isn’t yet supported by common voice assistants either. 

In terms of sentence interpretation (the messages), the technology has reached a quite good level in the past years, but we still don’t have human-like dialogues with the systems and, interestingly, users have been, so far, quite tolerant with that. 

When we interact with voice interfaces, we expect a certain response or trigger a certain function. We’re purposely interacting with the interface. Because we know its (still existing) limitations, we adapt our interactions to the capabilities of the interface. For instance, we tend to use shorter, more straight-to-the-point language with them.

The “robotic moment” we made reference to in the introduction may be playing a role here. We’re ready for the robots, fearful but ready, so we’re willing to be patient with them too. 

 
Voice assistants under scrutiny: the omnipresent privacy concerns

No doubt, there are privacy concerns around voice searches or interactions when using voice assistants. According to this article, 41% of voice users are concerned about issues like passive listening. 

Attached to our lives and the intimacy of our home, these devices record any noise around them, may that be conversations or any sensitive information. Some years ago, the most widely adopted voice assistant, Alexa, was accused of sending private conversations to a contact, because it thought the person called “Alexa” when using certain words that sounded similar. Accidentally—and hopefully not maliciously—the use of voice assistants can interfere with our private lives. 

People may not use a financial institution’s voice-based solution out of fear that someone may be listening. Or people may worry that banking apps always listen to them, even while they’re not using the voice features.

A second concern is the purposeful use of data to improve the systems and the technology supporting the voice assistant. The large AI organisations running these systems often record voice snippets to adjust and advance the algorithm. They are collected and sent to providers that check if the voice interactions were treated correctly. Users aren’t so aware of this, even though they agreed to this in the terms of service.

In financial services, banks have been first adopters of voice technology aiming at improving client engagement. This has taken them to explore how to improve client security.

Voice is a means of biometric authentication, as our eyes are. By launching voice-activated banking apps or enhancing existing apps’ security with features only activated with voice recognition, banks are upgrading security. 

Clients can perform financial transactions—for instance check their account balances, the transaction history, request information about new financial products and other details—by simply talking to a voice-enabled device. 

In Europe, this technology may still have a long way to go before being deployed. Banking security standards or the lack of regulation on the matter might forbid voice-enabled financial transactions initially.

However, voice technology can undoubtedly benefit from biometrics to increase security and go beyond passwords using your mum’s dog name or your favorite singer’s birth date.

Media scrutiny, clients’ acceptance…voice assistants’ future in financial services could emulate the smart glasses’ nosedive or it could be that of cryptocurrencies, rocky, yes, but steady.

 
Voice-based financial advice: could it truly add to the user experience?

Rules around personal financial advice are complex. To be not only efficient but also trustworthy, providing financial advice often requires empathy and wider social understanding.  

There is the technical side of it, but there is the human and social context too. Let’s recall the attempts to use wearables that enhance the human eye: they didn’t fail because of the technology in itself, but because they weren’t socially accepted. 

“Never say never ” is an important statement in the AI domain. The field of voice biometrics is already taking steps towards detecting emotion and stress in voice recordings. Actually, it has been already tested in the context of suicide hotlines around the world to support assessing the call’s urgency, Andreas shared with us. 

Judging on voice only is a challenging topic, even for humans. Sales via phone do not work as well as face-to-face, as the cues we pick up with our eyes are crucial. The social context is also a highly complex topic, argues Andreas, and AI research is not quite there yet. But it is feasible to expect a lot of progress on this in the coming years.

 
Picturing the future voice assistants in financial services

Coming with no surprise, there is this fact: providing financial advice via an online voice interface is technically more challenging than taking people to a website.

While voice-based technology tools are becoming more powerful, there are many “have to”: the voice interaction needs to be carefully modelled, the access to data has to be provided, the rules of the dialogue are required to be specified, and the system has to be appropriately tested for adverse results. Providers are consistently improving their tools for this purpose, but it requires considerable time and effort.

When asked about the future of voice assistants in financial services, Andreas was quite honest. 

A voice-only bank is nothing that we should foresee in the near future. Apart from the technical challenges, the scope might be too narrow. 

Offering access diversity to services is, however, an asset, so rather than pursuing the “best voice experience”, financial institutions want to aim for a “very good mobile, web, physical and voice experience” to capture as much of the market as possible.

 

Final thoughts

Machines have no friendship to offer, and yet we persist in the desire for conversation, companionship, and even communion with the inanimate. That comment was something we read and digested when preparing this article. And it’s a striking idea when you think about it. 

Maybe we all have, somewhere deep inside, the unspoken aspiration that one day we could interact with robots, voice assistants or any other AI-powered cyborg like we do with human beings. But as we are all unique, only time —and each one of us on his/her own—will tell. 

But for now, let’s keep watching what the upcoming generation of voice assistants will bring us. And because the handling of frustration is highly dependent on each person’s tolerance to it, let’s continue using our voice assistants and repeating, as we sometimes need to do, our commands.  

Each of us has a different robotic moment, after all. 

 
What we think
Andreas Braun, Director at PwC Luxembourg, Artificial Intelligence & Data Science
Andreas Braun, Director at PwC Luxembourg, Artificial Intelligence & Data Science

“Alexa, please, write an insightful quote for me.” This is something still out of reach for our voice assistants, but supporting our financial affairs is not. Well-implemented voice technology can increase security, convenience, and accessibility for clients, improving their overall experience and also helping financial institutions save costs already today.

Leave a Reply

Your email address will not be published. Required fields are marked *