[Podcast] Cognitive machine learning: “Ride the wave or get hit by the wave”

Katie Sadler bw
Posted: 04/13/2018

“The beauty of cognitive is that it is continuous learning—anytime the answer is right, it gets better and better,” said Roberto Mancone

CognitivePhoto by Vladimir Kudinov on Unsplash

“My background is not in digital. My background is not in IT. Actually, my background is 25 years of corporate banking and corporate finance,” said Roberto Mancone, former managing director of global head disruptive technologies and solutions at Deutsche Bank.

Taking over responsibility for financial products in Europe for Deutsche Bank in 2011, Mancone turned his attention to emerging disruptive technologies and consequently reinvented his career. “It surprised me that in two years I was able to recycle myself. Just by simply looking at the wave coming, I said to myself either ride the wave, or get hit by the wave,” he explained to AIIA Network Podcast host, Seth Adler.

With a background in process, Mancone was able to look at the organization and pinpoint challenges that not only Deutsche Bank had but the industry in general. “The cost of service is too high and the customer's attitude and habits rapidly change,” said Mancone.

“If that is happening in other industries, why can’t it happen with us in the financial industry?”

Mancone then went on to identify the technologies that now shake the way that the bank does business today. He achieved this by looking at what other industries had accomplished.” Industries like aviation, transportation, logistics, entertainment, traveling and hotels, these are also regulated environments but nevertheless, they went through major shifts, and somehow they were able to cope with the global reach of the customers, real-time reach of the customers. I said, if that is happening in other industries, why can’t it happen with us in the financial industry?

“My disruptive goal was to open a lot of doors outside of the bank. However, this may close a lot doors inside of the bank in the sense that you can imagine when you call a meeting and you are the head of disruptive, people think, ‘what is going to happen?’," he explained.

Moving solutions to AI and blockchain, Mancone revealed that operations were no longer in the design think era but were now in the implementation phase.

“We wanted to create an experience which was totally different from an unsatisfactory chatbot”

“We wanted to create a really seamless experience where the customers could be absolutely comfortable in talking their real natural language. So behind the machine learning, behind the extraction, we wanted to create an experience which was totally different from an unsatisfactory chatbot.

“But before this, we wanted to create is a culture in which employees were comfortable using something that isn't necessarily a human being before they even engage and pass this to the client,” said Mancone.

“You have to dissect the processes and define what the typical questions are? What are the typical topics? You then generate and train the artificial intelligence”

Before implementing cognitive technologies “you need to have knowledge of the processes, you need to have knowledge of the sales force, and you need to have knowledge of the products and the services that the customers request. You have to dissect the processes and define what the typical questions are? What are the typical topics? You need to define the intents. You then generate and train the artificial intelligence on that particular product and that particular service.”

“The beauty of cognitive is that it is continuous learning, so the system always trains and retrains and anytime the answer is wrong, the system learns that the answer is wrong. Anytime the answer is right, it gets better and better and better.”

Listen:

3 key takeaways from this week’s AIIA podcast

Unknown unknowns in the business model

“You need to put a lot of unknown unknowns in the business model. This creates a lot of uncertainty in choosing the right model at the right time to do the right projections. I have a solution for this… you can use artificial intelligence to do predictive analytics and business modeling for your business plan.”

Continuous positive experience

“The moment that you provide customers choice, then they are looking always for a better experience, convenience. And it's not just a matter of price, but a matter of comparison, benchmarking, a combination of solutions.”

A balancing act

“Implementing the technology  doesn't mean that you kill whatever you have or that you cannibalize whatever you have because you have the responsibility to the employees, the clients, and the shareholders to try to continue to generate profitability while you invest in something else. You need to make sure that investment is not going to jeopardize the return on the assets and internal investment. So, it's a balancing act that you need to maintain.”

You may also be interested in:

Banking on AI: BFSI in the age of artificial intelligence

AI 2020: the global state of intelligent enterprise

Katie Sadler bw
Posted: 04/13/2018

OUR BENEFACTORS