Business Automation:
IBM’s Business Automation is a complete and integrated platform for automating and digitizing work. It is designed to address insight challenges that that companies usually see in their business. Key benefits are:
- Increase employee productivity by automating repetitive and mundane human tasks.
- Drive down costs by leveraging AI to streamline processes and minimize business risk.
- Grow revenue from new and existing customers by delivering customized and automated client experiences at scale.
Use case example:
A customer decides to get refund as the product was not delivered by committed date. They fill-out an online form and they attach copy of invoice. Vendor makes a refund approval decision and, then notifies the customer and issues the refund. Here, there is no strait-through processing as it’s handled mostly by human. This is an issue! Goal here is to handle 80% of valid refund requests with no human interaction.
Let’s understand the current process:
First, customer submits a refund request which arrives in a general customer support inbox, opened by human who determines the priority, enters data into systems and routes the request to relevant person.
Next, a decision service evaluates the request on timeframe, loyalty status, amount and other data to determine the approval flow through people and systems which can be inconsistent and inefficient depending upon people and the system involved.
Last, all refund information is entered into the billing system before notifying the customer. Certainly, it’s not productive! Also, there may be chances of human mistakes.
Let’s see what happens with Robotic Process Automation (RPA):
IBM’s Robotic Process Automation (RPA) are added for accounting to enter data into billing system and customer support to select and fill out emails by template, to issue refund automatically. RPA can complete these manual tasks in few seconds instead of hours.
Let’s see what happens with Content Analysis:
Content analysis is added to process incoming documents, attached documents from the customer with different formats, such as invoice and receipts automatically.
IBM’s document processing system is able to classify incoming documents of various layouts and extracts the required data using machine learning to understand and learn about new labels and fields over time. By automatically capturing invoice data for refund requests that have an attached invoice, we can increase straight-through processing from ~30% to ~50% and reduce average processing time from ~4 days to ~3 days. It’s a huge improvement in customer’s company.
Let’s see what happens with intelligent refund chatbot:
With built-in natural language processing and AI, chatbot understands customer input, score confidence levels for each intent and learn over time. The chatbot also uses native RPA commands to query the company’s systems during conversation to validate and gather the required data and create the refund request in the workflow system, eliminating manual data entry. By providing automated real-time assistance to customers who have questions or need help, we can increase straight-through processing from ~50% to ~60% and reduce average processing time from ~3 days to ~2 days.
Finally, additional decision types can be more fully automated to maximize profitability by keeping its profitable customers happy through understanding risk and predicting each customer’s propensity to churn.
By combining decision services with machine learning based predictive analytics, for the customer’s propensity to churn and future lifetime revenue, the company adjusts their refund decisions to be more while at the same time providing significantly better and faster customer service, including a reduction in average time from request to refund to less than a day. It increases straight-through-processing from ~60% to ~80%, achieving our goal of 80% straight-through processing.
