Intelligent process automation for operations
Robotic process automation (RPA) is an exciting development for banking and financial services companies. This technology shifts repetitive operations, such as checking, inputting, searching and collating information, from humans to software tools. By interacting with multiple applications at the software presentation layer or user interface and processes, RPA mimics the steps that humans would take, reducing expenses and errors in the process.
Lower costs and fewer mistakes are only the beginning, though, of automation’s potential benefits. Intelligent process automation, or IPA, builds on RPA by both freeing people to take on higher value work and enabling them to work smarter using process data analysis. Results of our recent survey of senior business and technology decision-makers from across banking and financial services sectors suggest IPA will be welcome in many institutions. Nearly half of respondents reported at least 10% revenue growth driven by process-aligned analytics over the past year. And, nearly three out of four banks expect to see similar growth in the next three to five years.
IPA tools use artificial intelligence and machine learning to deliver accurate, time-sensitive results. By integrating IPA, banks can potentially reduce the time, effort and cost of operations and processing, while taking advantage of meaningful insights from automation-generated data. We expect IPA to deliver performance improvements including: reduced error rates; better management of repeatable tasks; improved standardization of process flow; frictionless straight-through processing; and less reliance on multiple systems.
Banks can implement IPA into their day-to-day operations by understanding business needs, identifying processes that can especially benefit from IPA, and making process improvements or changes. Three activities in particular provide good examples of this potential: wealth management, investment banking and mortgage services.
Investment advisors work directly with their high-net-worth clients, forming trusted, one-on-one relationships. At the same time, banks are seeking to expand their investment services to other customers, including investors not interested in such relationships and millennials seeking economical investment advisory options.
IPA captures data on investor behavior, risk profile, life stages and decisions and feeds it to an artificial intelligence (AI) engine. The AI engine analyzes this information in combination with market data to generate and continually improve investment recommendations for the advisor’s use. IPA can also increase investment advisor efficiency. Advisors can expand their client base significantly by automating much of the analysis they currently perform.
For investment banking sales and trading professionals, the business day typically has three distinct parts: assemble and analyze voluminous data and formulate trading strategies before the bell; adapt strategies to economic events during trading hours in communication with clients; and assess the performance of sales trader recommendations after the close.
IPA can help sales traders throughout their busy day. A self-learning AI engine incorporates multiple data sources and applies rules and preferences to recommend strategies. Economic events during the day are automatically captured and analyzed. Based on the analysis, recommended strategies are revised and provided to sales traders instantly along with a list of potentially affected clients. At the end of the day, performance data on sales traders’ recommendations arrive on their digital devices in an intuitive graphical format.
IPA can generate similar benefits in the mortgage process, which typically includes long waits for applicant approval; dozens of manual application handoffs; errors and missing documents; balky digital channels; and high lender costs. IPA can help make the process more bearable for everyone. The AI engine, for example, can improve and expedite loan approval decisions by expanding the data factored into applicant profiles. Information such as social media profiles and activity can be incorporated with credit scores. Deep analytical insights at relevant process steps can minimize delays and errors during loan processing and servicing. And queuing of complex cases for higher-level approach can speed processing times.
Some of the characteristics to consider when identifying which processes are candidates for IPA include:
Completion time. Analysis of the average time required for various processes to complete can help determine which ones are leading candidates for automation.
Cost. Targeting data-intensive processes and combining data for automation can potentially provide greater savings.
Error rate. Knowing the error rate at various process steps can help determine the impact of corrective action either during the process or downstream.
Data adequacy. Determining if data is sufficient at each process step or additional data is needed will be critical to completely automating a process.
Degree of intelligence (artificial or human). Intelligence at each step can be AI, human or hybrid, such as an AI recommender with a human approver.
Optimal process. Analysis can determine if the process is making the optimal decision or there is opportunity to improve decision making by combining processes or algorithm-based straight through processing.