Using AI to Automate & Optimize Commercial Lines Underwriting

Artificial Intelligence - May 26 2020

According to a recent poll we conducted, underwriters are spending 35 to 50% of their time on core processing with only 25% of their time spent on selling and broker engagement. This equates to underwriters spending half of their working day on tasks like manual policy checking, re-keying data into multiple systems, and reading submissions, rather than on strategic judgment work or nurturing broker and customer relationships.

In a data-rich industry like insurance – probably one of the most data intense industries in the world – 80% of the data is trapped in data lakes or silos across disparate systems, resulting in issues relating to data cleanliness, data hygiene, data degradation and data mismatched across different datasets.

Artificial Intelligence (AI) has the ability to help remedy data issues by taking unstructured data that is stored in a data lake or some type of repository and creating a structured output that allows underwriters and underwriting administrators to have greater access to data (like never before) to make informed decisions and better risk selections.

AI and Humans Working Together

There is a lot of debate about AI replacing humans, however, AI and humans or skilled knowledge workers each have their own strengths. We often hear concerns that AI is taking work away from humans. However, the reality is that AI has no ability to perform judgement work like a human. Skilled knowledge workers or humans far outperform machines when it comes to judgment work.

On the other hand, AI can read and extract data from unstructured insurance documents hundreds of times faster than a human. Commercial insurers who harness the power of AI to extract, interpret, contextually understand and resolve hundreds of data points in seconds, give their underwriting teams access to a lot more data from a variety of sources enabling them to make better informed decisions and accelerate the underwriting process.  

Not All AI is the Same

In a webinar we recently hosted, we touched on the fact that there are many different types of AI including Natural Language Processing (NLP) and Machine Learning (ML). At the core of AI is machine learning which is the ability to teach a machine or computer how to read and within the machine learning branch of AI, there is deep learning, supervised, semi-supervised and unsupervised reinforcement learning.

One example of AI is natural language processing or NLP, is the ability for a computer to read and contextually understand text or language like a human. In our day-to-day lives we all have experience with and probably even take for granted Siri, Amazon Echo, and Google Earth which are all good examples of AI question n’ answer systems. You ask it a question such as “What’s the weather?” and it gives you an answer. This is an example of speech-to-text and text-to-speech. The machine is understanding the human language and converting it into text and generating a human-like voice so we can understand what it is saying when it speaks back to us.

Another example of NLP is content extraction classification. It is the ability to read all the entities or nouns in a document. Things like people, places, numbers, etc., and classify them such as limits, policyholder, premiums, etc. An AI solution, like Chisel AI, that is purpose-built for insurance can extract and contextually understand more than 500 insurance terms or data points such as the insurer name, the insured name, insured city, insured country, etc. This specific type of NLP is called Named Entity Recognition (NER), which is the ability to automatically recognize, extract and contextually understand entities or nouns. NER that is designed for insurance is capable of extracting hundreds of data points from unstructured insurance documents at superhuman speeds and with more accuracy than a human. The system can recognize data points in insurance documents like policies, binders, quotes, submissions, applications, statements of value, loss run reports, vehicle schedules, etc., in digital formats such as email, Word, Excel, and PDF.

Unlike RPA or Expert Systems, one of the oldest forms of artificial intelligence, NLP systems continuously learn and get better with experience and do not rely on well-defined rulesets or decision tree logic like “if this, then that” that require ongoing upkeep and maintenance. RPA is good for solving a specific business problem or issue, whereas, NLP and ML systems that speak and understand “insurance” can provide cognitive processing at any point within the underwriting process with little or no reliance on a human. AI truly allows insurance companies to plan and reimagine processes.

Revised Ron Glozman Blog Quote #1  Seven Business Reasons to Use AI

It does not matter if you are a carrier, a broker, a reinsurer or an agency or what line of business you write, the insurance industry is very paper-laden. There are a lot of processes that require manual human intervention that are creating barriers to growth and missed opportunities.

Revised Ron Glozman Blog Quote #2

There are many ways that AI can automate and streamline underwriting and brokering processes like data extraction, policy checking, submission intake, and submission prioritization to name just a few.

AI-powered underwriting delivers operational efficiency and resiliency enabling insurance companies to quickly capitalize on new opportunities, enter new markets and drive data-driven risk decisions.

Here are seven ways AI benefits commercial insurance carriers and brokers:

  1. Reads and extracts data locked in unstructured commercial insurance documents at superhuman speeds giving underwriters access to greater insights for better risk assessment and pricing
  2. Eliminates data silos and improves data hygiene across multiple back office systems
  3. Expands underwriting capacity by 50% enabling insurance companies to double their business without adding staff
  4. Reduces costly E&O risk and the time spent on manually reviewing policies through AI-assisted policy checking
  5. Automates and streamlines high-volume, mind-numbing, repetitive manual underwriting tasks
  6. Accelerates quote to bind
  7. Provides operational resilience and efficiency with 24/7 availability (AI does not take holidays!)

For more insights on how your organization can benefit from using AI to automate your underwriting processes, how AI works, and top considerations for selecting the right AI solution for your organization, watch our on-demand webinar “Using AI to Automate Underwriting.”

Webinar Image 1200 X 628 (2)

Additional AI for Commercial Insurance Resources:

Browse different topics

Recent Posts