Digital Banking

Why Context Makes Virtual Assistants Smart Digital Bankers

AI and ML-powered services can help fight banking fraud. However, to layer the tech into viable solutions to assist customers with banking, robo-advisors need to truly become “smart,” intuitive — and conversational. In a new podcast, Brighterion CEO Akli Adjaoute tells Karen Webster that “smart agents” tech predicts what consumers need and want — by putting everything in context, word by word.

In the melding of tech and banking services, the goal is an ambitious one: to create a private banker in digital form  a chatting, knowledgeable persona that interacts with the consumer, using twin high-tech engines of machine learning (ML) and artificial intelligence (AI).

The ideal digital banker knows (seemingly magically so) about the consumer’s individual situation  whether they’re starting a family, for example, and might be mulling buying a home. Its suggestions about mortgage financing are helpful, even intuitive, and take the stress out of remembering to plan each step of a financial journey that is life-long and can get side-tracked amid the busyness of daily routine.

Easier said than done.

In a podcast interview with Karen Webster, Dr. Akli Adjaoute, CEO of Brighterion, a Mastercard company, pointed out that the legacy machine learning and AI currently powering digital assistants like Alexa, Siri and Cortana lack both adaptability and the ability to determine context.

More on context in a minute. As a refresher, and as has been seen by PYMNTS, it should be noted that the data is available for AI platforms to adopt and adapt, to bring robo-advisors to consumers, along with the “wow” factor of anticipating what users want and need. The fact remains, though, that successfully mining this data, and using it to bring self-learning to those assistants, has yet to be perfected.

Things are getting there, it seems. Adjaoute told Webster that advancement is still a ways off in terms of replacing humans with AI, but harnessing Smart-Agents technology allows financial institutions (FIs) to get a 360-degree view of the consumer and their activities to make sure they are both secure and well-served.

The security? Well, we’ve already seen how unsupervised machine learning and AI can reduce reliance on rules-based fraud monitoring and cut false-positives drastically for banks. As Adjaoute recounted in a past interview, that means less friction in the consumer experience, and a boon to both banks and the banked. However, when it comes to a robust services model, flexible and marked by personalization (i.e., as illustrated above, where the virtual assistant anticipates needs before they are voiced), some real fine-tuning is necessary, according to Adjaoute.

“The only way you can have a really intelligent assistant,” he said, “is if the virtual assistant is able to understand something without any pre-programmed language.” In other words, they must go beyond stored and structured data; they must rely on “other words” — literally.

Context In Context

Adjaoute gave an example. The word “apple” can have a series of meanings: It can refer to a computer (c’mon, it is 2018 and it’s probably the first thing one thinks about), a fruit or even travel — as in a journey to “the Big Apple.” True AI, he said, allows the system using the technology to find the true meaning of how a word is being used, even when a word can have multiple meanings. It’s all about context.

He pointed to the principle behind Smart-Agents, technology that creates a virtual tag to entities — in this case, words — and builds profiles as these words are used. There are as many smart agents as there are words, each one tied, say, to every word in the English dictionary, or to the German or French. As these smart agents interact with one another, they learn on the fly (so to speak, and no pun intended) to put things in context.

Adjaoute maintained that it is context that matters when seeking to put the “human” factor into the services business, in this case banking services. No longer are rules universally applied to consumer interactions with FIs. Instead, the smart agents update their own profiles through each interaction and action, learning how those consumers act as individuals.

Overall, he said, “you have to get the technology to understand the fundamentals  what’s the meaning and what’s the relationship  because words are like poetry when you speak well.” Every word, he added, expands the scope and insight into what is being discussed, and what outcome may ultimately be desired.

“We as humans use context to extract the main idea, and then we zoom down to the last sentence or the main idea,” he said. “We are really trying to reproduce the way our [minds understand] things and to put that with the Smart-Agent[s] technology,” he told Webster.

Against this backdrop, he said, there will most certainly be an increased appetite to use smart agents and smart technology to grow the service-related revenues of banks, with an eye on client retention.

“So, rather than bombarding people with the offer or request that no one cares about … if you concentrate on what is likely to interest them today and [financial products] they will use, technology has no limits on what it can provide,” he told Webster. “I do believe that artificial intelligence is the way that will allow any business to onboard the customer, serve them and deliver the absolute perfect service to them at the right time with the right products, which will be beneficial for both” FI and consumer.

——————————–

LATEST INSIGHTS:

Our data and analytics team has developed a number of creative methodologies and frameworks that measure and benchmark the innovation that’s reshaping the payments and commerce ecosystem. Check out the latest PYMNTS Digital Drive Report 

TRENDING RIGHT NOW

To Top