A skilled underwriter’s time is precious, but low-skilled and low-value communications work constantly intrudes on their working day. Fortunately, Communications Mining can extract the critical insight from digital messages, freeing underwriters to focus their skills where they’re needed most.
In our recent AI Pioneers interview with Ocado Chief Data Scientist Gabriel Straub, Gabriel described ‘cognitive energy’ as the most precious resource a business has. As humans, we’re mostly single-threaded beings - though a lucky few could be described as multi-threaded. Regardless, all of us have a finite cognitive capacity - that is, there’s a limit to the number of things or tasks that we can devote our full attention and resources to at any one time.
This is nowhere more true than in insurance. Years of training and experience go into creating a skilled underwriter. However, they can only focus on so many different tasks in a given day. Their specialist knowledge of complex underwriting criteria is hard to replicate, so simply ‘hiring more’ isn’t a solution to growing case volumes. To increase capacity and maintain their cognitive energy, insurers need to help their existing underwriters do more work - or more precisely, more valuable work and fewer chores.
Reducing the number of manual, repetitive tasks and intellectual waste underwriters must deal with is the best way to improve productivity. Automation has been used to great effect in driving faster and smarter underwriting through automated responses and data-driven decision making. Yet its true potential remains untapped. A new approach is needed to drive worthwhile automation at the greatest source of process inefficiency - email.
Email: The source of intellectual waste
An underwriter’s work is based in communications. Across high-volume, high-complexity lines like Personal, Commercial and Claims, underwriters spend much of their time in email. While focused on conversing with brokers and claimants, the tasks are often transactional in nature. They’re constantly exchanging information between parties or following up for a timely response. This is essential work, but what often isn’t realised is just how inefficient it is.
On taking action to improve process efficiency, specialist insurer Hiscox discovered that 97% of emails sent by brokers were routed to the wrong person on its underwriting team. Indeed, simple tasks like email triage and requesting more information take up an unsustainable amount of underwriters’ time. Considering 30-40% of their day already consists of administrative tasks, this leaves little time for valuable work - like risk assessment and policy underwriting - that really makes use of their skills and training.
Manual email processing is generally low-skilled and repetitive. It represents true intellectual waste for skilled underwriters, increasing latency, handling times and costs. However, automation to date has been of little value here. Email communication depends on the processing of conversational, natural language. Humans do this innately, but it has been very difficult to teach machines how to replicate this skill. Most automation tools require structured, machine-readable data to produce accurate decisions and useful outcomes. Until that data becomes available, underwriters will continue to spend much of their day trapped in Outlook.
Communications Mining: Insight and automation
Email is an unavoidable cost of doing business. However, insurers can greatly reduce that cost and give precious time back to their underwriting teams through Communications Mining.
Communications Mining is an emerging class of enterprise software that extracts value from unstructured conversational data - including emails. It uses natural language processing (NLP) and machine learning to understand masses of business conversations at scale, and converts them into structured data that can then be used for insight and automation.
The application of NLP to communications has seen mixed results in previous years. But recent advancements in model accuracy and training have made NLP-based solutions safer, more reliable and valuable than ever. New models now routinely outperform humans in language understanding and reading comprehension. NLP deployment is now a safe, impactful and cost-effective option, provided the insurer does not attempt to build its own solution from scratch.
Communications Mining gives underwriting teams unprecedented, real-time insight into their email-based processes. Furthermore, it enables straight-through processing in workflows once impossible to automate. With minimal training, today’s machine learning models can rapidly learn when key information is missing from an email, or understand when the wrong recipient has been chosen. Automated workflows can easily be set up to resolve these issues, taking manual repetitive work out of the underwriter’s hands and freeing them to focus on their most complex and valuable tasks.
Communications Mining is a critical stepping stone, scaling automation into once opaque unstructured comms channels. However, it’s benefits aren’t limited to automation. Communications Mining also gives underwriters a 360° overview over inbound communications, regardless of their seniority or technical skill. Teams can analyse what the most common drivers of work are, and which requests take the longest time to process. With these insights, they can plan, resource and automate much more effectively, enhancing productivity in one of the most time constrained parts of the business.
To see the benefits of Communications Mining in action, read about how Re:infer helped Hiscox achieve comms process automation for a better broker and underwriter experience.
