A few years ago ING Bank saw the power that AI can bring to various parts of their organization. In particular, visionary technology leaders at ING saw the power that AI could bring to bond trading, allowing them to make faster and sharper pricing decisions. Sharing his insights on an AI Today podcast interview, Santiago Braje, Head of Credit Trading at ING, explained more about how ING sees AI in the context of bond trading, and the vision behind its Katana platform.
Predictive analytics pattern to help bonds traders
Named after the ultra-sharp Japanese Samurai sword, Katana uses the pattern of predictive analytics to help traders decide what price to quote bonds when buying and selling for their clients, based on both historic and real-time data. Katana began as an augmented intelligence platform, and it was first used as a tool to assist traders working specifically inside ING. The problem Katana addresses for market makers is how to respond to requests from clients, especially in bond marketing..
In the market, the investors need to get in contact with banks to buy bonds. In the past it was common for them to call the banks to get the process moving, but as time and technology has progressed investors have been going through electronic methods more frequently. As electronic trades have become more common, the market has seen a shift from larger, infrequent trades, to smaller but more frequent trades. In turn, the market has become more competitive and investors look to get multiple quotes at the same time. The problem now is how can banks respond to these requests. Katana is designed to make this decision making process sharper and faster.
Katana is designed to combining data visualization with artificial intelligence algorithms. It brings all relevant information to one screen for the trader to see what they need. This information includes historical trends, current trends, and projected data trends, providing context for the trader so that they can make decisions quickly and more efficiently.
As any new innovative project, Katana faced some challenges as it took time to get off the ground. One of the largest challenges was that in a large, established company such as ING, when developing new technologies to be used, it can be a challenge to work around already existing infrastructure, organization, and complex systems. Another challenge was working around the interactions between already existing technology and systems. In addition, the next major challenge was convincing users of buying into the new technology. They knew they would need to build trust and relevance to the traders if adoption was to happen. The final challenge was developing internal talent around the new innovation, including finding people who understood what needed to happen as well as had the correct skills because this area of trade is incredibly competitive.
At the time of the interview, Braje stated that the program had only been in use for a few months. However at that time the results already showed three major successes. The first was that speed in trading increased immensely. This was an expected result and the most obvious benefit gathered by the program. The next big triumph was opportunity price increase. Pricing and accuracy increased for the users of Katana. And the final major result was that some users saw increases of three to four times more in trade by using Katana than they would have seen before. The project was initially rolled out with a small amount of users,but it shows many potential benefits as the program grows and gathers more steam.
Of interest to those who follow the use and implementation of AI programs is how they are developed and what sources help to drive them forward. Braje shared how exactly this was accomplished for Katana. Some of the technology sources for the program came from open source technology and others were developed completely from scratch. The benefit of open source coding is that there is a huge community of users and developers to help catch things and problem solve. The program was also designed as a modular extendable platform so that the program could eventually have things added on and designed later. The company has since expressed a desire to expand the AI-based technology to be used more broadly within the Netherlands-based financial services company
Internal department to move forward innovation
ING is unique as a bank because it has an entire department dedicated to internal innovation and even provides the opportunity for those in the company to tap into a venture fund for innovation. Braje points out that ING is really an innovation company with a banking license. The company has set up an innovation office and funding with a board that makes decisions on what innovations can be greenlit, and helps provide support for the whole process of developing and implementing these projects. The fund has been around since 2016 and the Katana project was one of the first projects to be funded.
AI will help us get to a place of making more sophisticated decision making from trade to developing ways to sift through data. As organizations like ING continue to find value in AI and make investments accordingly, more and more departments will create tools like Katana for help their employees extract more value and hidden insights from current data.
