Information Extraction for Commercial Insurance Underwriting

Artificial Intelligence, Natural Language Processing - July 24 2019

Standardization of data continues to be an ongoing problem in commercial insurance with many different formats, data fields, and poor data quality hampering underwriters from effectively assessing and pricing risk.

With broker submissions taking on many different forms including emails, PDFs, word documents, ACORD forms, spreadsheets, requests through agency or broker management systems and various other artifacts, underwriters are forced to scour through documents to find, retrieve and analyze the data required to assess risk and make decisions. These submission intake challenges continue to impede insurers’ productivity and hinder straight-through processing.

For example, one email submission may contain eighteen attachments including a schedule of vehicles, loss run reports, existing policy, and other supplemental forms. And in extreme cases, broker documents can be ten to hundreds to even thousands of pages long. Today, an underwriter is required to manually search each of these documents one by one to retrieve the data needed to evaluate if the business meets the risk appetite. It’s a tedious, painstaking process that slows down response times.

Using technologies like Natural Language Processing (NLP) and Artificial Intelligence (AI), real-time data extraction can automate and streamline commercial insurance processes such as policy checking, submission triage, and submission prioritization. By automatically reading and interpreting key data points in insurance documents faster than a human and auto-populating core underwriting systems, AI-powered data extraction solutions allow underwriters to focus on critical thinking, complex reasoning, and strategic account management.

According to Cognizant, any process or technology improvements in submission intake will instantaneously allow underwriters to focus on better risk selection, pricing, and customer excellence, and not worry about the operational inefficiencies and other activities that add little value to the organization. 

Reveal Business Insights and Automate Underwriting Processes with AI

Using AI, commercial insurers can free trapped information from unstructured and structured insurance data sources such as policies, quotes, submissions, applications, binders, loss run reports, statements of value, etc. Natural Language Processing and AI solutions that harness Named Entity Recognition (NER) can automatically recognize hundreds of insurance-specific data points or nouns such as name, city, premiums, endorsements, such as limits, premiums, deductibles, types of coverage, exclusions, endorsements, etc. faster than a human and with significantly greater accuracy.

With integration to a centralized submission inbox or portal, AI can extract data at a granular level from applications and submissions as they come in. Data can be parsed and interpreted from tables and other content blocks, and then auto-populated in core underwriting systems such as quoting systems and rating engines through an API integration. Intelligent solutions that leverage natural language processing have the cognitive capabilities to read, interpret and understand data associations within the document like humans can. The ability to read the data is not based on the location of the data in the document – that is, AI goes beyond recognizing patterns and like-for-like similarities and understands the language and context.

With exposure to business insights previously hidden in documents, data lakes and data stores, underwriters can make better-informed decisions to quickly and efficiently triage submissions and identify the best business to write, saving them enormous amounts of time and effort.

According to a report by Accenture, leading insurers that use intelligent solutions to reinvent the customer experience and to drive human-machine collaboration are achieving returns in excess of 10 times their investment in technology.

For insurers to achieve their growth goals, it’s critical to identify the submissions that meet the chosen risk appetite quickly to eliminate submission bottlenecks and missed revenue opportunities. With only a small percentage of submissions actually quoted and even fewer that are bound, by augmenting skilled knowledge workers with digital employees, insurers can automate and streamline underwriting processes, expand their underwriting capacity, quote more submissions, and allow skilled staff to spend more time on strategic point of sale work.

With a purpose-built AI solution for commercial insurance that understands insurance jargon and includes insurance workflows out of the box, commercial insurers can automate the application intake process by setting up auto-decline rules that trigger automated emails to brokers, notifying them if their submission is declined. Simultaneously, they can trigger auto-route rules to route submissions that fit the risk appetite to specific folders for specialists to review and write the business. By setting up business rules that can filter submissions based on industry code, years in business, the number of claims, annual revenue, this allows insurers to automatically assess if the submission meets their risk appetite and supports their book of business.

These are just some of the results commercial insurance companies have experienced:

  • Extract and interpret 500 data points in one second, compared to a knowledge worker who can extract only 15 to 50 data points in 30 minutes
  • 50% increase in underwriting capacity without adding staff
  • 2X increase in quoting capacity with no additional headcount
  • Process and respond to thousands of submissions in a day

AI Augmentation Boosts Productivity

Extracting data points from unstructured and structured insurance documents faster than a human and sharing that data to downstream systems has significant business benefits to insurers as they strive to achieve growth goals while controlling costs. Gartner predicts that by 2021 AI augmentation will recover 6.2 billion hours of worker productivity.

As commercial insurers look to optimize the digital customer experience, automating and streamlining the underwriting process will allow underwriters to make decisions in minutes as opposed to days and weeks, improve distribution communication, and enable the firm to better manage risk.

Accelerating quoting times and improving the frequency of communication with brokers can earn insurers the status “carrier of choice” as brokers can count on them for fast, accurate quotes. By winning over brokers, insurers can secure more future submissions.  

Want to Learn More? Tune in to Our Webinar

Join the live webinar hosted by ACORD: Emerging Trends in AI, Data Extraction & Underwriting (Wednesday, September 4, 1:00 PM ET). You'll hear expert insights from Ajay Agrawal, Co-Author of Prediction Machines: The Simple Economics of Artificial Intelligence and Ron Glozman, CEO of Chisel AI. Everyone who signs up for the webinar will be entered in a draw to win one of ten copies of the best-selling book Prediction Machines

SAVE YOUR SEAT 

 

Browse different topics

Recent Posts