Know My Company
What hype do you see around the applications being used by insurance companies today that are leveraging AI and ML?
There is certainly a lot of hype around AI–ML today. However, when we work with insurance companies, they are bringing us very critical and specific use cases that they hope will help them avoid expensive processes and solutions and error-prone activities. The use cases we’re seeing are all requiring multiple customer touchpoints.
How are insurance companies leveraging data for the purposes of enhancing the customer experience?
Let us consider a specific use-case of Email triaging. In this example, the insurance company has historical data in the form of several millions of emails. The company wants to bring that data to a solution provider who can support the right combination of structured and various degrees of unstructured data. I say unstructured because many of those emails may contain attachments. It wants to use that content to train the AI-ML models so that incoming emails can be triaged, classified appropriately and have their data extracted. Those emails need to be routed to the right subject matter. What’s more, the company wants to have the ability to auto-respond to customers requiring specific user responses to continue the engagement and bring a claim to its conclusion.
From a customer perspective, it’s important for the insurance company to send an immediate context-based acknowledgment email. This acknowledgment could include requesting further pertinent information to expedite the processing of a claim. Most insurance companies understand that the ability to rapidly engage with a customer has a direct correlation to high customer satisfaction.
What types of applications are being created using AI that Insurance companies are deploying today?
Some applications are used to handle the FNOL (First Notice of Loss) process; some are used for email triaging (responding to and the intelligent routing of emails); and some are used to automate the underwriting process, which may include smart decision assistance based on underwriting criteria and historical data.
How are insurance companies applying cognitive technologies for process automation in quote intake, claims, training and customer service use cases?
Many insurance companies are in the process of evaluating solution providers that provide AI-ML capabilities. The decision-making process is often based on testing a small amount of data. Once an insurance company has selected an AI-ML-based solution provider, the next step is for the vendor to determine where the solution fits into the insurance company’s technology stack. From there, the insurance company will import its domain-specific data into the vendor solution for AI-ML training. They’ll then test the system by feeding it various input information.
For quote intake, insurance companies gather historical data, which usually amounts to lots of emails with attachments of different formats, and use it to feed and train the models. They will then inject sample emails to test the system for accurate data extractions. In some cases, they also integrate with the backend systems.
When it comes to claims processing, insurance companies today are looking for solutions that combine both the conversational aspect as well as the process automation aspects. This is because claims processing involves extracting data sets out of claim data and then using the data to engage with customers to retrieve further information. Before this technology was available, the claims process took between one and three weeks to close a claim. Today’s conversational interfaces have reduced that time to just a few hours. Once all of the information is gathered, it is then processed and automatically integrated with the backend system for the next steps.
In the case of AI-ML training, insurance companies are using a vast array of materials that are retrieved by traditional systems, allowing training teams to do high-level topic searches that are then used for training models that will be used for automation. Insurance companies are now deploying cognitive technologies to do deep searches on archived data, and the technology supports free-text queries to retrieve results with high significance. Instead of manually traversing document hierarchies, cognitive technologies allow for flat searches across all documents. For this to work, insurance companies must feed large volumes of training materials to the AI-ML systems where they continually index and learn in order to provide accurate search results.
Customer service is a great use case for this technology. For example, in the case of incoming emails that includes customer requests, the emails are triaged, auto-responses may be triggered, and the messages are then routed to the right subject matter experts. It’s done this way because the AI-ML models involved in triaging are already pre-trained with historical emails. This leads the software to classify incoming emails with a high degree of accuracy. Automating customer requests allows a company to get back to customers quickly and retrieve further information if needed. Using this approach, customer requests are handled efficiently, which results in huge cost savings and leaves customers with the favorable impression that their requests were taken seriously.
How can insurance companies best impact customer experience while leveraging AI-powered automation?
Customer experience along with automation is the most successful equation. For automation to be successful during a customer interaction, the platform should combine a conversational-AI interface with process workflow automation. When the two are fused together, the handshake that occurs between related subsystems should be able to meet business demands.
Thank you, Henry! That was fun and hope to see you back on AiThority soon.
Focused on guiding and empowering all teams including engineering primarily to deliver fitting solutions for our customers. Making sure the entire solutions and underlying architecture are well aligned towards business goals, short-term and long-term and constantly making sure all aspects and principles of products are kept intact. To build with clarity on what matters, with simplicity to keep the focus on primary goals of the solution, to put in place structures and processes so that they aid us in accelerating our work rather than becoming a burden, scaling both product & teams, advising every one in the team as and when required, keeping informed and staying aligned with CEO and the board are the ongoing work I am constantly involved in. Never short of challenges and never short of accomplishments amidst failures. Quite interesting.