The future of AI and ML in banking with Kony’s Bill Bodin
- Kony has recently rolled out two new products that incorporate AI in different ways.
- CTO Bill Bodin joins us on the podcast to talk shop on AI and ML in banking applications.
Banks have been talking about leveraging artificial intelligence and machine learning for years. for the most part, it was mostly talk. Now, real use cases of AI and ML are taking hold inside financial institutions. First and foremost, these technologies are being applied to the customer experience.
Bill Bodin is CTO at Kony. He’s been working with these technologies for years, beginning earlier in his career at IBM, where he was the CTO of mobile. At Kony, he’s recently rolled out conversational interfaces via an acquisition and a larger banking technology development platform that gives mobile and web developers the AI-delivered support of ace programmers built into Kony’s Quantum platform.
We talk about how banks and credit unions are beginning to introduce AI/ML to their customers and the use cases it most makes sense to start with. We drill down into how these technologies are being used to help customers make better financial decisions and how they’re specifically being used in the lending application process. Lastly, we talk about inherent bias in systems like these and what Kony and the rest of the fintech industry is doing to combat them.
I spent 26 years at IBM. I ended as the CTO for mobile as a distinguished engineer. I explored many things over my career there, like operating systems. I created an advanced technology lab to study the internet of things before there really was an IOT.
Especially toward my latter years at IBM, I had exposure to machine learning and artificial intelligence. I worked closely with the Watson team to create linkages into mobile and other solutions.
Our latest acquisition was a firm called Pivotus. They had created a product called Engage, which if focused on bringing customers and bankers together in real time conversations. We make it relevant through a layer of AI, Natural Language Processing, Sentiment Analysis, and Machine Learning.
It creates a stickier relationship between banks and their customers, and also makes it more meaningful, as well. Being able to data mine large data sets in real time and deliver the results of that into a human-human chat interaction — it can be really meaningful.
We probably showed our earliest chatbot interactions at Finovate 2017. Even that, integrated personal wealth management, opened the aperture as conversational user interfaces are expanding. It wrappered in retail APIs to allow a customer of a bank to know if he could afford to buy a particular product. It crosses over many systems, including retailers’ APIs, and at the end of the day, the bank was able to suggest deals that made sense to the consumer.
Eliminating biases from new financial technology
The industry is beginning to see it as an issue. From our perspective, we’re really focused on fairness. When you consider how fast algorithms can process data, it becomes hyper important to eliminate bias from the logic. In most cases, AI, NLP, ML systems are trained by humans. While machines function without bias, the rules and initial training can carry bias.
We make sure that bias is eliminated from the training process and the tools we use. Google has been very proactive in blocking gender-based pronouns from one of its high-level composition features in a step to remove bias from the equation. AI incorporates image recognition, a challenge in recognizing people from all backgrounds — based on the training models that have been used.
I think we’re turning the corner on bias — both gender and racial. Tools, algorithms and training methodologies are becoming more fair.
We’re super excited because we just shipped our newest product, Kony Quantum. It’s a platform to enable developers to create low-code applications. It’s also our first release to include AI as an integral part of the development platform.
Now, you can have chat features within the application that actually respond and help you as if there’s an expert developer sitting next to you. It can also look at what you’re doing and make suggestions. It can say things like ‘that looks like an authentication for an application or at least the start of one — here are 30 examples of forms that live in our marketplace that all have backend connectivity to OAuth and other forms of authentication. Would you like to import those directly into your application and save a lot of time?’