Communications work has become a real burden in underwriting, but insurers can automate a great deal of it with natural language processing, according to Ed Challis, chief executive of Re:infer
An insurer’s most precious asset is its underwriting team. Underwriters bring in new business, make money and manage risk. Any time they aren’t focused on assessing risks or underwriting policies is time and talent wasted. Unfortunately, this happens all too often.
Communications work is a necessary evil. Underwriters need to communicate with brokers and customers to get the insight and information they need.
Junior underwriters will spend much of their time on policy amendments, to free up the time of their more experienced colleagues.
However, this email-based comms work is highly inefficient. People make mistakes, brokers leave out important information, or send their messages to the wrong team member.
Underwriters are left to pick up the pieces, performing menial tasks like email triage, which make little use of their precious time and abilities.
The automation gap
Inefficient processes not only waste time, they create an operational bottleneck. If time-poor underwriters are struggling to maintain their inboxes, they aren’t closing new business or focusing their efforts on retaining existing clients. The higher the volume of communications, the worse the problem becomes.
It’s crucial that insurers target automation efforts at email inboxes, helping underwriting teams understand, route and automate email requests correctly the first time, every time.
However, email and unstructured communications data presents a serious obstacle.
Familiar automation tools, like robotic process automation (RPA), need structured data to execute processes. The natural language contained in emails can’t be so easily understood or actioned by this technology.
That’s why skilled underwriters continue to spend so much time simply reading and forwarding emails. This limits trained experts to the work of glorified clerks.
Conversational data intelligence
To win back time for underwriters, insurers will have to consider new approaches to automation.
Fortunately, natural language processing (NLP) has achieved considerable advances in recent years and is now capable of reliably understanding and actioning email.
Through a process called active learning, today’s NLP models can be trained to immediately identify the nature of a request, who it’s meant for and if any information is missing. These details are then converted into structured data, which is shared with automation tools for downstream processing.
An underwriting team would benefit greatly from a NLP solution that can understand and process every inbound email.
If a broker has forgotten to share necessary information in their message, the solution can request it automatically. If the email has been sent to the wrong person, it will be forwarded immediately - no input needed from the underwriter.
When insurers and underwriters have this unprecedented insight into their comms channels, we call it conversational data intelligence. This new category of technology is helping free underwriters from the demands of manual email processing, enabling them to focus their skills where they’re needed most.
