AI To Impact

PODCAST: Making AI Real

Episode 2: AI enabled Risk Management for FS powered by BRIDGEi2i Watchtower

Listening time: 13 minutes

AI enabled Risk Management for FS powered by BRIDGEi2i Watchtower

Today the Chief Risk Officers(CROs) struggle with the critical task of monitoring and assessing key risks in real time and firefight to mitigate any critical issues that arise. An Automated Risk Surveillance system to adequately alert them to credit risks and respond quickly to dynamic market scenarios is the need of the hour to help them navigate through the complex ecosystem.BRIDGEi2i‘s AI-accelerator Watchtower, is fundamentally designed to fit into multiple use-cases of Financial Services, such as, tracking response and customer conversions for marketing and cross-sell, delinquency, NPAs and losses for credit risk, and so on. It helps companies identify opportunities for proactive risk intervention.

Monica: Hi, everyone. Welcome to another edition of our podcast series, AI to Impact, by BRIDGEi2i, a podcast on making AI real. My name is Monica Gupta and I lead design and deployment of contextualized AI solutions for our clients. powered by BRIDGEi2i’s proprietary accelerators.

These solutions help drive both functional and pan-enterprise transformation by making AI real.One of the crucial asks in enterprises today is the need for agile business surveillance, executives want real-time business performance monitoring with the ability to get contextual alerts and deep dive analysis. BRIDGEi2i’s proprietary accelerator Watchtower delivers AI-powered business surveillance with automated unexpected business event detection, anomaly alerts, and advanced root cause analysis. Business users across various verticals and domains have been using this to reduce their time to actionability, thus preventing impact to their businesses.

In the previous podcast, we heard Predeep Ghimrey, an expert in the CPG industry who talked about the use cases of Watchtower in the CPG domain. Today, we have with us Ashwini Agrawal who leads financial services practice at BRIDGEi2i. Ashwini has extensive experience in designing solutions for clients using advanced ML techniques to harness information from different forms of data.

Hello Ashwini, it’s so good to have you with us.

Ashwini: I am doing very well. Thank you for having me, Monica.

Monica: Thank you for joining in, Ashwini. So Ashwini, in today’s data driven world, information is getting generated from various directions and deriving meaningful insights remains a constant ask for businesses. Although BI systems have evolved to help business leaders with these insights, we feel that it’s still not able to provide those actionable insights. You have been working for over a decade in financial services and risk management programs. Do you see a similar challenge among your clients as well?

Ashwini: Hmm, see, BI systems of yesteryears provided good visibility on key metrics, but they were descriptive at best and had clear challenges as we started asking for more. Too many reports and metrics, but low clarity on what really needs attention. Almost no highlight of why exactly something is happening and lack of clarity on what needs to be prioritized and acted upon.

It’s very reactive, not proactive right? So, someone needs to go to a dashboard and identify any red flags. And then a lot of the analysis needs to be done offline, which significantly impacts speed to decisions. See every decision maker–right from the CXOs to portfolio managers, campaign owners and business operation managers need to keep a tab on some key metrics that matter to them and are in pursuit of understanding two aspects primarily.

One, what can I do better to improve those key objective metrics and two, can I proactively identify any potential adverse impact to those metrics? Because if I can identify those adverse impact, I can then potentially mitigate them before they become large. These decision makers, I would imagine would love to have a virtual analyst who monitors these key metrics through multiple layers of the dashboards and brings up signals and insights on those two aspects, proactively and timely.

Monica: Hey, so as you rightly mentioned, and I would hundred percent agree to that; given complex business hierarchies, fast changing business scenarios and time-consuming insights, there seems definitely a need for an automated business surveillance system which can help business leaders act in an agile and responsive way to stay ahead in the game?

Ashwini: Yeah, absolutely. From my experience with financial services, an automated business surveillance system finds use across a spectrum of marketing, cross-sell, credit and operational risk, collections, investment risk, onboarding of customers, digital journeys, et cetera. And then an AI-enabled business surveillance accelerator like the Watchtower constantly monitors signals for all KPIs across all granular dimensions and its ability to proactively identify and alert on any anomaly or deviation that needs attention, truly enables faster and better decisions. Talk about Watchtower right, it’s just so fundamental in design that it can fit into multiple use-cases. If I talk about some Financial Services ones, tracking response and customer conversions for marketing and cross-sell, delinquency, NPAs and losses for credit risk, you know, and audit and compliance, quality control metrics for operational risk, tracking return on investment for investment management portfolios, top off percentages for digital and online journeys, et cetera, et cetera. So use cases can really be fit in by all business and function leaders because all of them are tracking some key KPIs that matter to them and the fundamental design of the Watchtower enables them to use it to their advantage.

Monica: Looks like you have Watchtower use cases for a lot of business problems in financial services. We talk about marketing risks, onboarding and investment. Can you elaborate on any one of them?

Ashwini: Sure, Let’s pick the use case of credit risk watchtower which we deployed for the CRO of a US-based MNC bank and resolved. The objective being, bubble up any deviations in the trends of key metrics that need attention of the CRO.

For example, delinquency 30+, delinquency 90+ and NPA trends, tracking across MSAs and regions and product lines, and identify if it is being driven by a specific risk score band, debt to income band or customer segments, or impacted by specific policy decisions or macroeconomic factors like house price index, unemployment, et cetera.

Some noteworthy aspects were not just being able to look at time trends of these metrics for a specific MSA or portfolio segment, but also being able to compare those MSAs across peer groups. For example, see if a specific MSA is consistent under-performer compared to its peer MSAs. Another noteworthy one would being able to calculate impact of the root causes as well, so the CRO could truly prioritize his or her actions and decisions across acquisition, collections or policy actions based on the expected impact. The CRO doesn’t just get the root causes for the deviations or anomalies that he or she sees, but also has an understanding of the impact based on which he or she can take decisions and actions.

Monica: Well, that’s interesting. Not only is it saving time in doing all this analysis and alerting of the root causes, but also bringing visibility at the higher levels. So, in addition to alerts, and root causes and the impact on investment metrics, what else do users like in your deployed solution?

Ashwini: Hmm, sure. That’s an interesting question.Monica. See, Watchtower is a truly intelligent solution. We can call it AI or BI, it uses carefully orchestrated algorithms and graph-based structures under the hood to provide the business user a seamless solution, helping drive prescriptive and timely action. So, the users, they don’t have to worry about what’s under the hood, they can use it in a seamless fashion. Plus, it’s a collaborative platform so users from different personas, responsible for the different root causes can come together, collaborate, coordinate and act so that different root causes can be addressed by different personas in the ecosystem and work in a collaborative fashion to fully mitigate the impact. And I will just add one additional aspect that it even helps with predictive simulations to understand the impact of potential business decisions or actions that a leader is taking before the leader wants to freeze on them. So, before I say here is the action I want on the whole of the portfolio, I can truly assess the impact of my decisions. For example, impact of my acquisition decisions, losses decision for the CRO.

Monica: Interesting. So, it is not just an anomaly detection tool. It’s a lot more than that. Okay. So you talked about Watchtower capabilities, which include proactively tracking the metrics and identifying any potential adverse impact on them. So, I’m curious to know. Across the use cases you talked about what kind of benefits your clients have seen?

Ashwini: Sure. Let me take a couple of examples. We deployed it for a US Fortune 100 bank as a risk watchtower to provide proactive alerts on key delinquency metrics, both off-net and on-net along with insights on root causes and impact.The estimated reduction in NPS by about 8%.

The second example would be where we deployed the Watchtower as a fraud decision engine, to provide proactive alerts for loan applications that need investigation along with the potential reason or root cause to enable better and faster fraud investigations as well. Playback what we saw there was that about 20% better fraud capture rates than the system that was already existing. So in this case the client already had a fraud detection system in place, and we had a better impact of 20% by using the Watchtower.

Monica: Okay. That’s really impressive. Seems like autonomous business surveillance is an emerging area in financial services as well. However, these are testing times for everyone. So, do you continue to see traction during these stress times as well?

Ashwini: Oh, Absolutely. If anything, it being more traction, Monica. The CXOs today wanted a closer watch on the portfolio…the portfolio metrics and observe any patterns on something going wrong, or some opportunity being left untapped. For example, it is even more important today for portfolio managers to keep a close tab on credit risks and investment risks and understand the root causes fast enough to take actions on high-impact areas with current age, that’s going through for last six months…clients opting for digital ways, data capture is better now. For example, across various onboarding and underwriting steps, the data capture has become better and the clients want to use it to understand prescriptive insights for drop-off and rejection reasons and fine-tune their onboarding journeys, models, strategy, thresholds, et cetera. So, I would say yes, very relevant in current times. And if anything, only becoming more relevant.

Monica: Okay. Thank you, Ashwini. That brings us to the end of this wonderful interaction. We heard about some very important use cases in the financial services domain, where Watchtower can really drive transformation.Thank you, Ashwini for your time today. It was indeed a great discussion. I’m sure our listeners would find this relevant.

Ashwini: I enjoyed the discussion too, Monica. My pleasure.

Monica: Thank you so much for listening to this episode of the AI to impact podcast, that was such an interesting conversation with Ashwini, financial services practice lead at BRIDGEi2i Analytics. It has surely sparked a lot of thoughts, ideas, and actions. Do subscribe to our channel as we continue to bring to you the latest in all things AI, through interactions with experts and thought leaders in our upcoming episodes till then goodbye, stay home and stay safe.

With new data formats, evolving data ecosystems and AI technologies becoming more mainstream, the intersection of Data Science and AI-powered interaction systems present a whole new gamut of opportunities for the digital enterprise. In our latest AI to Impact podcast series: Making AI Real, we will be interacting with several thought leaders, BRIDGEi2i Business Heads, Domain Experts, and reputed AI and analytics leaders to learn more about endeavours and innovations happening to make AI real for enterprises and gain tangible benefits to stay ahead in the business. Stay tuned to know more!

Meet the Speakers

Monica Gupta

Monica Gupta, Manager, AI Actions

Monica Gupta is Manager, Data Engineering and AI Accelerators at BRIDGEi2i, and leads the Design and Deployment of contextualized AI solutions for clients at BRIDGEi2i. Bringing in the techno-functional perspective, she handles roadmap, solutions and delivery of AI accelerators as part of her current responsibilities. She has led some sizable Digital Transformation programs in the areas of Customer Experience, Supply Chain, Sales and Digital Marketing across CPG, Enterprise Tech and Consumer Tech companies. Monica holds a Master’s degree in Finance from Delhi University.

Ashwini Agrawal

Ashwini Agrawal, Director – Financial Services

Ashwini Agrawal leads the Financial Services practice at BRIDGEi2i. He has extensive experience in designing solutions for clients using advanced ML techniques to harness information from different forms of data to power AI actions across customer acquisition and growth, risk management, and customer service. Ashwini has previously worked with HSBC, Capital One, and [24]7 Inc in different roles. He holds a Masters from IIT Kharagpur in Physics (5 yr. Integrated). Ashwini was named among AIM’s 40Under40: Leading Data Scientists in India 2019.