Digital Transformation in Commercial Lending is proceeding apace with banks rolling out initiatives in multiple areas like Re-imagining customer experience, Empowering Relationship Managers, and adopting Intelligent Automation techniques. Leveraging data analytics across the lending life cycle can be a key enabler in each of these areas.

Leading Banks are climbing the analytics maturity ladder: from Descriptive Analytics (‘What has Happened?’) to Predictive Analytics (‘What Could Happen?’) to Prescriptive Analytics (‘What Should Be Done?’). Progress along this scale represents significant advancement in the sophistication and maturity of their data modelling capabilities, and can enable them to reap significant benefits across all the key transformation areas described earlier.

Surveying the Data Landscape:

We begin with a survey of the types and sources of data that banks can leverage across the lending value chain:

• Structured Data: includes data residing in application databases, and also data delivered in structured formats like XML & JSON through API calls
• Unstructured Data: unformatted & semi-formatted data residing in documents such as credit memos prepared by loan underwriters, meeting memos prepared by Relationship managers, news articles, analyst reports & social media feeds

• Internal Data: available within the bank’s application environment from systems like CRM, loan origination, risk rating, loan servicing, etc
• External Data: available from data feeds from external providers like Bloomberg, corporate filings with stock exchanges & government databases, news reports published on the internet

• Quantitative Data: Financial Statement data, financial ratios, macroeconomic indicators
• Qualitative Data: Commentary recorded by underwriters in credit memos, meeting notes captured by relationship managers, research reports published by industry analysts

Key Industry Initiatives

With this background, we now take a look at key initiatives banks are working on at each stage of the analytics maturity scale:

Descriptive Analytics – ‘What Has Happened?’: The use cases in this section focus on leveraging analytical insights to provide insights on historical data

1. Customer 360 View: Aggregating 360 degree views of their customers to enable Relationship Managers & other personnel to access relevant information about their customers. Information captured in these may include: Relationship overview, key management profiles, key financial metrics, peer comparisons, upcoming events,  covenant compliance status, ongoing deals, deal pricing simulators, cross sell opportunities, alerts & notifications
2. Portfolio 360 View: Aggregating customer 360 degree views for the set of clients managed by a Relationship Manager or underwriter or along domensions such as region, country or industry segment leads to a Portfolio 360 degree View, which may include: Portfolio composition, aggregate exposure across different product lines, portfolio metrics, etc. This view can also present different views of the portfolio by slicing and dicing across dimensions like industry, risk rating, geographic mix, etc.
3. Account Insights for Customers: Providing insights to customers on their lending relationship such as data on approved limits and utilization, due dates & repayment schedules for loans, account balances, covenant compliance status, peer comparison, etc.

Predictive Analytics – ‘What Could Happen?’: This section describes use cases to leverage existing & historical data to predict future outcomes

1. Cash Flow Forecasting: By analyzing the cash flow patterns of clients, banks can predict when the business is likely to face a cash flow shortfall. A timely offer of additional credit limits at this point can help the cleint to meet such working capital gaps. The best sources of data to perform this analysis are the client’s accounting packages and Enterprise Resource Planning systems, and banks are offering incentives to their clients to share this data with them. Alternatively, account statements can be analyzed to perform such analysis. Models developed to forecast cash flows can also be leveraged to predict future covenant compliance.
2. Environment Monitoring: News articles, analyst commentary & research feeds can be curated for Relationship Managers & underwriters, to help them to keep themselves up-to-date on developments relevant to their clients, and the industry segments and geographies in which they operate. Sentiment scoring algorithms can be applied to determine whether a particular news item has a positive or negative impact. A composite sentiment score (weighted by recency) can be computed to obtain a single quantitative value to represent the current sentiment on any client, industry or geography. 
3. Early Warning Signals: Proactively predicting financial stress based on lead indicators based on Modelling of financial, behavioral, industry, macroeconomic indicators. This analysis can also identify different attributes which have predictive power at different time horizons (eg: 6 months, 12 months, 24 months, etc).

Prescriptive Analytics – ‘What Should be Done?’: We now list use cases where data can be leveraged to deliver actionable insights to bank stakeholders

1. Next Best Conversation: Empower Relationship Managers with contextually relevant suggestions on topics to discuss with their customers. Examples include providing an industry viewpoint, cross sell opportunities, offering credit facilities, etc
2. Loan pricing & deal structure assistance: Relationship Managers can be empowered with tools to model various scenarios for pricing & structuring a loan so as to be able to provide more realistic advice to their customers on-the-go. This tool could suggest optimal deal pricing terms such as interest rate margins, fees, collateral requirements etc to meet the bank’s pricing hurdle rates. It could also be used to suggest pricing terms offered to other peer borrowers for comparison.
3. Underwriter Assistance: Banks can mine credit memos prepared by underwriters across the bank to provide recommendations on optimal terms and conditions such as covenants, financial benchmarks, monitoring conditions, etc. These suggestions can help to share best practices and assist underwriters in reducing turnaround times.

Conclusion

Data analytics techniques like machine learning and artificial intelligence promise to enable significant leaps in delivering on the use cases listed above by enabling banks to process large volumes of heterogenous data efficiently. Banks must strive to develop a full service capability across the entire maturity scale to derive optimum returns.