The Game Changers: AI in Banking

Recent advances in AI have thrown the field wide open for a broader set of business leaders, who are realizing its potential to reorient their organizations and transform data-driven insights into real-world solutions.

Throughout 2020, AI analytics has helped major banks stay connected to their customers – especially during the dramatic rise in use of online channels to manage finances.  As leading banks applied a range of successful strategies to navigate the new normal, AI analytics also opened greenfield opportunities to raise profits, and offered lessons in agility and ethical responsibility.

The arrival of COVID-19 – combined with the specter of Big-Tech looking to gain a stronger FS foothold – introduced a new way of being, forcing banks to develop better ways to serve customers. Most turned to AI.  Many turned to AI analytics to provide the high levels of functionality, data visualization and presentation that meet customers’ rising expectations.

In retail banking, game-changing AI-driven benefits range from analytics-backed personalized offers (eg loans), to regular portfolio overview (savings/investment recommendations) and daily financial health snapshots (with bill-payment reminders).  For SME customers, they include customized lending (loan offers generated by projected company cashflow), smooth inventory management, analytics-backed sourcing of suppliers and buyers, and beyond-banking services (automated tax returns, ready for signing).

According to McKinsey Global’s The State of AI in 2020 survey, half of cross-industry respondents said their organizations adopted AI – in at least one function – to increase revenue and cut costs.  

At SparkBeyond, we’ve seen a steady rise in AI analytics-adoption by the big banks, who have used it to generate at-scale hyper-personalization, enhance omnichannel experiences and embed rapid innovation cycles.

Two distinct client case studies reveal how market leaders have used AI to match the speed and flexibility of Fintechs – driving impact within 10-12 weeks and steering their businesses through the pandemic.

To keep pace with radical shifts in consumer behavior – and uptick in credit requests – AI analytics can be used to generate rapid, precision data on probability of default, loss given default, exposure at default and credit conversion factors.

The resulting risk assessments identify micro-segments of previously un-lendable applicants:  gig economy workers, students and entrepreneurs.  This enables retail banks to widen their new customer base, and capitalize on the market share potential of the ‘non-traditional customer/thin-file’.

In short, bank the unbanked.

(Past efforts to accurately risk-model the underserved segment have failed.  A top global bank recently spent 2 years – and approximately $200 million – with no tangible benefits.)

Greenfield Digital-First Bank

‍A top European client launched a greenfield digital-first bank to acquire 500,000 newcustomers from an emerging “solopreneurs” segment – a complex, thin-file of clients whose unconventional behavior makes them difficult to bank.

‍Result:

‍Within weeks, SparkBeyond’s platform connected six (previously siloed) internal and external datasets, revealing 50 million patterns driving risk, while our external (real-world) datasets generated additional, highly valuable, and previously hidden, insights. Our model – with an eight-fold improvement in accuracy compared to the baseline model – powers the data backbone that now serves the client’s new thin-file customer base.

Until now, data science teams – operating manually – have struggled to leverage up-to-date insights at speed, and uncover the hidden patterns and correlations that impact cross-sell/upsell metrics.  Today’s AI analytics evaluate millions of ideas per minute, revealing multi-layered insights within hours, accurately forecasting customer activity trends and customer life events which go beyond normal routine.  Predictions are used to generate clear graphic rendering of customer segmentations for targeting, and devise innovative products for new micro-segments – boosting wallet share, cementing customer loyalty and reducing churn.

‍AI Analytics for Marketing & Sales

‍A leading Swiss financial services company was looking to deploy AI across its marketing and sales divisions to map customer journeys across a range of its insurance products.  SparkBeyond discovered and assessed more than 100 actionable drivers behind new customer acquisition, as well as existing up-sell and cross-sell behaviors.  

‍Result:

‍‍Using 1,000 data fields and combining them with GDPR-compliant external datasets, SparkBeyond’s Discovery Platform uncovered hidden insights which led to a seven-fold increase in returns from targeted campaigns.

This initially generated more than 4 million Swiss francs of potential bottom-line, annual value (at an ROI of 15x in 5 months). Today, the Discovery Platform is embedded into the company’s analytics stack and used across a number of business units by analysts, data scientists and business stakeholders.