Future Finance: A Look into AI’s Role

Ali Curi: Markets Conversation is a new ION podcast where we discuss topics of importance to capital market participants with product owners, subject matter experts, and industry leaders.

Riccardo Bernini: The adoption of AI has been pushed by the explosion of Big Data. Consider that nowadays, every one of us in our interaction with other human being is producing a lot of data every minute, the interaction with our smartphones or the PC, so a huge amount of data can be collected.

Ali Curi: Hi everyone and welcome to Markets Conversation. I’m Ali Curi. On today’s episode, we’re discussing the role of artificial intelligence in financial services.

Ali Curi: Our guest today is Riccardo Benini from ION’s LIST Professional Services, and he will help us understand how the financial services industry is undergoing a major transformation due to the emergence of artificial intelligence or AI. AI is being applied in a variety of ways in financial services to automate everyday tasks, provide more sophisticated financial advice, detect fraud, score credit and the handling of personalized services and back office operations.

Ali Curi: Our guest today is Riccardo Benini from ION’s LIST Professional Services, and he will help us understand how the financial services industry is undergoing a major transformation due to the emergence of artificial intelligence or AI. AI is being applied in a variety of ways in financial services to automate everyday tasks, provide more sophisticated financial advice, detect fraud, score credit and the handling of personalized services and back office operations.

Ali Curi: As AI continues to advance, financial organizations must embrace the technology in order to gain a competitive edge and create new opportunities for the company’s growth. Let’s get started.

Ali Curi: Riccardo Bernini, welcome to the podcast.

Riccardo Bernini: Ali, thank you for the opportunity.

Ali Curi: Let’s start with some background on our topic today. Let’s get a general overview of AI in Financial Services. So, Artificial intelligence in banks. Why now?

Riccardo Bernini: AI, I think has become an essential part of technology in banks, insurance, and in generally in financial sector. And since it allows to, improve the quality of the offer that provide to the customer. If we look internally at the traditional processes of this kind of company, I think that AI and Machine Learning allow to speed up this process to make them more efficient and in particular, to reduce human errors.

Riccardo Bernini: AI, I think has become an essential part of technology in banks, insurance, and in generally in financial sector. And since it allows to, improve the quality of the offer that provide to the customer. If we look internally at the traditional processes of this kind of company, I think that AI and Machine Learning allow to speed up this process to make them more efficient and in particular, to reduce human errors.

Riccardo Bernini: So this is why the success of this kind of models because they allow to reduce cost and in the meanwhile to provide better services to the customers.

Ali Curi: Let’s dig a little deeper. Let’s discuss some of the upsides. What are some of the benefits of AI in financial services?

Riccardo Bernini: I would tell you one of the benefit that I think, and I think it is the possibility for a bank to make better informed investment.

Riccardo Bernini: I would tell you one of the benefit that I think, and I think it is the possibility for a bank to make better informed investment.

Riccardo Bernini: I’m not thinking only to financial investment, but I’m thinking to strategical investment. So consider a manager of the bank. Today, he has new technology that allows him to understand better what is going in the bank, exploiting the big data that the bank is collecting, as I said, to take more informed decision and to make more better strategical investment.

Ali Curi: And what would you say are some of the challenges of using AI in finance?

Riccardo Bernini: Oh, there are several challenges, but today I will tell you three, four points. And I think the first challenge is data quality because consider that in data science there is a saying and it is, “garbage in, garbage out.”

Riccardo Bernini: What does it mean? It means that if you could have a best model in the world, but if your data are not cleaned are not curated and well maintained, the model will not perform because the results are strictly related to the input. So you have to set up disaster recovery in order to guarantee the continuity of the acquisition of your data and to set up policies that allow us to, as I said maintained and clean the data.

Riccardo Bernini: What does it mean? It means that if you could have a best model in the world, but if your data are not cleaned are not curated and well maintained, the model will not perform because the results are strictly related to the input. So you have to set up disaster recovery in order to guarantee the continuity of the acquisition of your data and to set up policies that allow us to, as I said maintained and clean the data.

Riccardo Bernini: This is the, I think, the most important point for the challenges and related to the data there is also another point, that is dimensionality reduction. Consider that when you manage Big Data there is a lot of noises in the data. There is a lot of rumors. So you have to be able to distinguish what is the signal to what is rumors.

Riccardo Bernini: And this is done performing what is called feature selection analysis. And you need to have a good data scientists that are able to understand, which is the key driver to feed your model and provide good results in output. And the last point that I want to tell you is the black boxes, “black box effect.” Today, there is still a lot of people and especially who are not a technical people, but it comes from the business that considers these model as black boxes because it’s difficult to understand the result, to read the metrics that allows you to deeply understand the result and the decision that the model advise you.

Riccardo Bernini: And this is done performing what is called feature selection analysis. And you need to have a good data scientists that are able to understand, which is the key driver to feed your model and provide good results in output. And the last point that I want to tell you is the black boxes, “black box effect.” Today, there is still a lot of people and especially who are not a technical people, but it comes from the business that considers these model as black boxes because it’s difficult to understand the result, to read the metrics that allows you to deeply understand the result and the decision that the model advise you.

Riccardo Bernini: I think this is one of the challenges and consider that there is a new branch of AI that is called Explainable AI, that as the aim to make the result understandable. These are the three key drivers, the three key challenges.

Riccardo Bernini: I think this is one of the challenges and consider that there is a new branch of AI that is called Explainable AI, that as the aim to make the result understandable. These are the three key drivers, the three key challenges.

Ali Curi: Now, Riccardo, at this point it’s pretty safe to say that AI is disrupting finance and banking.

Ali Curi: What would you say are some of the drivers of this AI disruption?

Riccardo Bernini: Okay. The AI, I think has been pushed the adoption of AI by several factors, but I would say the first is the explosion of Big Data. Consider that nowadays, everyone of us in our interaction with other human being is producing a lot of data every minute the interaction with our smartphones or the PC. So a huge amount of data can be collected from the world and in order to be able to manage this data and to extract right information, the standard, traditional statistical model is no longer suitable.

So this has pushed the adoption of Machine Learning and AI, which are able to process this huge amount of data and do this deal and extract information.

Riccardo Bernini: Okay. The AI, I think has been pushed the adoption of AI by several factors, but I would say the first is the explosion of Big Data. Consider that nowadays, everyone of us in our interaction with other human being is producing a lot of data every minute the interaction with our smartphones or the PC. So a huge amount of data can be collected from the world and in order to be able to manage this data and to extract right information, the standard, traditional statistical model is no longer suitable.

So this has pushed the adoption of Machine Learning and AI, which are able to process this huge amount of data and do this deal and extract information.

Riccardo Bernini: And the second point that pushes the adoption of Machine Learning AI, is the availability of infrastructure. Consider data storage, cloud computing, GPU. These are all factors that allows the adoption of this kind of model. And also in software area, today you have a lot of Open Source packages that allows you to use the model that the model are already implemented, you only have to pick a package and use it.

If we look at the business side, I think that what is pushing the adoption is a regulatory requirement. Because banks are highly regulated as we know, and regulation require a lot of report and with strictly cutoff timing. And, in order to fulfill this requirement, you have to think, automation.

Riccardo Bernini: And the second point that pushes the adoption of Machine Learning AI, is the availability of infrastructure. Consider data storage, cloud computing, GPU. These are all factors that allows the adoption of this kind of model. And also in software area, today you have a lot of Open Source packages that allows you to use the model that the model are already implemented, you only have to pick a package and use it.

If we look at the business side, I think that what is pushing the adoption is a regulatory requirement. Because banks are highly regulated as we know, and regulation require a lot of report and with strictly cutoff timing. And, in order to fulfill this requirement, you have to think, automation.

Riccardo Bernini: And automation is provided by this model. And these, I think these are the key drivers for the adoption and disruption of the AI in the sector.

ION Advertisement: This episode is brought to you by ION. At ION, our clear derivatives solutions, automate your complete trade lifecycle and deliver actionable insights whenever and wherever you need them.

ION Advertisement: We offer execution and order management, post trade processing and a complete front-to-back business solution. To learn more, visit us at iongroup.com/markets or email us at [email protected].

Ali Curi: Let’s explore what are some examples of some practical applications? We’ve talked a little bit about the drivers that are pushing forward, helping the banking sector apply these methodologies.

Ali Curi: Let’s explore what are some examples of some practical applications? We’ve talked a little bit about the drivers that are pushing forward, helping the banking sector apply these methodologies.

Ali Curi: What are some examples of practical applications in the banking sector that are either being used now or could be implemented soon?

Riccardo Bernini: Okay, there are several application, but I would like to start with two application that we already implemented in our products in my company. And the first is trade execution in the trade execution area.

Riccardo Bernini: And the second is market surveillance. In the straight execution area, the problem is to execute big order on the market. And executing this order is crucial to avoid market impact. So you have to split the order and decide which is the quantity for each child order and which is the timing for the releasing of this order.

Riccardo Bernini: And this is a very complex optimization problem and AI comes into play because is able to find the best strategy for releasing these orders in the market and allows the trader to achieve his goal. That could be to make more P&L or pay less fees to the market.

Riccardo Bernini: And this is a very complex optimization problem and AI comes into play because is able to find the best strategy for releasing these orders in the market and allows the trader to achieve his goal. That could be to make more P&L or pay less fees to the market.

Riccardo Bernini: In the market surveillance area. It’s an area which the regulators, require that the banks monitors the activity of their customer on the market in order to find out if there is some suspicious behavior. And at least some I would say some market abuse. Typical market abuses are for instance, insider trading or market manipulation.

Riccardo Bernini: The point is that when a bank adopt a software that allows to monitor the activity of their customer, there are a lot of false positives. So the banks typically put a lot of effort in terms of human resources to analyze the alerts and to flag them if they are to be closed. A false positive have to be signaled to the regulator.

Riccardo Bernini: The point is that when a bank adopt a software that allows to monitor the activity of their customer, there are a lot of false positives. So the banks typically put a lot of effort in terms of human resources to analyze the alerts and to flag them if they are to be closed. A false positive have to be signaled to the regulator.

Riccardo Bernini: And here, machine learning can help the banks in particular, the compliance officer, to analyze these alerts and to automatically flag them as false positive or to be signaled. Okay, so these are the two application that we already have in our products, but there are many other area, and I could list you this application and for instance chatbots, as we know in our home banking and other website, fraud detection, predictive analytics, or credit scoring for mortgage assignment.

Riccardo Bernini: All this kind of application are already matured and AI and machine learning is used in this kind of application.

Ali Curi: So what would you say is next for AI in financial services? Because what you’ve shared so far seems to hold great promise for transforming the industry. There are some areas which clearly focus on the backend.

Ali Curi: So what would you say is next for AI in financial services? Because what you’ve shared so far seems to hold great promise for transforming the industry. There are some areas which clearly focus on the backend.

Ali Curi: Some other applications focus on the customer experience. What are some areas where we can expect to see continued development and innovation?

Riccardo Bernini: For sure I would say that natural language processing is the most promising area. And consider that NLP has a lot of fields of application, even in the financial sector. For instance, they are the base, NLP is the base for chatbots, is the base for a recommendation system, is the base for fraud detection and prevention. So there are a lot of area where NLP can be applied and we saw at the end of the 2022 what the ChatGPT has been able to do and surprise the world with its capabilities.

Riccardo Bernini: For sure I would say that natural language processing is the most promising area. And consider that NLP has a lot of fields of application, even in the financial sector. For instance, they are the base, NLP is the base for chatbots, is the base for a recommendation system, is the base for fraud detection and prevention. So there are a lot of area where NLP can be applied and we saw at the end of the 2022 what the ChatGPT has been able to do and surprise the world with its capabilities.

Riccardo Bernini: So ChatGPT is obviously based on NLP. And I think that today we are just at the starting point for this technology. So in the future, NLP could help the banks, the financial sector, to take competitive advantage if a bank can exploit this technology.

Ali Curi: So you mentioned ChatGPT and most of the applications that we’ve discussed so far have, are for business and or commercial use, but consumers are also getting into, jumping on the AI bandwagon. You mentioned, like I said, ChatGPT, but there’s also Google, Bard and Bing ai. What are your thoughts around these popular, more consumer- facing AIs?

Riccardo Bernini: I think that it could help a more comfortable interfaces between the banks and the customer. So more close relationship between the banks and the customer, because consider just 10 years ago or 20 years ago, when you have a problem with your banks, you have to go to the local office of the bank to talk with the people. Today you have own banking and I think that in the future you will have the chatbots that really could help you in any of your problem with the banks. So I think the way in which banks and their customer will communicate and will interact each other in the future is changing, will change a lot.

Riccardo Bernini: I think that it could help a more comfortable interfaces between the banks and the customer. So more close relationship between the banks and the customer, because consider just 10 years ago or 20 years ago, when you have a problem with your banks, you have to go to the local office of the bank to talk with the people. Today you have own banking and I think that in the future you will have the chatbots that really could help you in any of your problem with the banks. So I think the way in which banks and their customer will communicate and will interact each other in the future is changing, will change a lot.

Ali Curi: Riccardo, what is the one big thing that you hope listeners will take away from this episode?

Riccardo Bernini: I think that, as I told you, I work in this field, I’ve been working in this field from many years and I still see in our customer, we are a software company, so we need to sell our products to our customer. But I still see some they are still worried about using this new methodologies. My hope is that in the future, a new culture about this methodology could grow so that a large scale adoption could be take place and I hope that this podcast could help a little. I don’t know.

Riccardo Bernini: I think that, as I told you, I work in this field, I’ve been working in this field from many years and I still see in our customer, we are a software company, so we need to sell our products to our customer. But I still see some they are still worried about using this new methodologies. My hope is that in the future, a new culture about this methodology could grow so that a large scale adoption could be take place and I hope that this podcast could help a little. I don’t know.

Ali Curi: So let’s talk a little bit about career advice and productivity. What is some career advice you wish you had heard earlier in your career?

Riccardo Bernini: If I had to talk to a young people, I for sure would say that the first point is keep learning. Because every day there, there are new papers, new models coming out.

Riccardo Bernini: If I had to talk to a young people, I for sure would say that the first point is keep learning. Because every day there, there are new papers, new models coming out.

Riccardo Bernini: So you have to keep learning, stay up to date. And this is what I would say to, to myself if I was at the starting point of this career. The second thing I would say is to build a network. So attend in industry events, take part in online communities, connect with other professionals because this is one way and for and maybe the most important way to stay up to date to talk with other people.

Riccardo Bernini: And because this world is constantly evolving, you need to have great networking with others. And last thing is to be patient because if you don’t find your ideal job don’t be worried, because if you keep learning, keep experimenting, you can build your portfolio or skills and sooner or later you will find your ideal job.

Riccardo Bernini: And because this world is constantly evolving, you need to have great networking with others. And last thing is to be patient because if you don’t find your ideal job don’t be worried, because if you keep learning, keep experimenting, you can build your portfolio or skills and sooner or later you will find your ideal job.

Ali Curi: Now, Ricardo, you’re a busy person juggling multiple responsibilities. How do you get things done? Can you share with us what is your go-to productivity hack?

Riccardo Bernini: Okay. As I told you, we are a software company. We are used to building application. We every day face with the problem of building application and make them usable to our customer.

Riccardo Bernini: But I think that for AI application, there are three more points that you have to consider if you want to build your product. The first point is that you have to deal with the big amount of data. So you have to project your database in order to have a big storage and fast access to the data. So you have to query the database in a quick way.

Riccardo Bernini: But I think that for AI application, there are three more points that you have to consider if you want to build your product. The first point is that you have to deal with the big amount of data. So you have to project your database in order to have a big storage and fast access to the data. So you have to query the database in

a quick way.

Riccardo Bernini: The second point is time consumption, because when you train a model, often it takes a lot of effort considering that, just as an example, I told you before that, we have a product in the trade execution and consider that to train that model, it took about one week with a lot of market data.

Riccardo Bernini: So you have to consider to adopt cloud computing or GPU when you build application in this area. And the last thing I would say is related to the black box effect because in order to make the results of the model understandable to business people, so they are not technical people, so you have to think your graphical user interfaces to be able to explain in a quick and easy to understand the result of the model, and I think this is the most challenging point of the three that I told you.

Riccardo Bernini: So you have to consider to adopt cloud computing or GPU when you build application in this area. And the last thing I would say is related to the black box effect because in order to make the results of the model understandable to business people, so they are not technical people, so you have to think your graphical user interfaces to be able to explain in a quick and easy to understand the result of the model, and I think this is the most challenging point of the three that I told you.

Ali Curi: Riccardo Bernini, thank you so much for joining us today. It’s been a pleasure having you on the podcast. I hope you visit us again.

Riccardo Bernini: You are welcome, and thank you to you for the opportunity and it was pleasure.

Ali Curi: And that’s our episode for today. You can follow ION Markets on Twitter and LinkedIn. Thank you for joining us. I’m Ali Curi.

Until next time.