Insurance Expert Systems

Decision Support for Insurance

Although commercial policy administration systems can offer some decision support, they often come into play only after key early-stage decisions have been taken; decisions about whether the risk is within the general appetite of the company and what products, endorsements and discounts the company wishes to make available to a particular customer. Policy admin systems are often product-specific and tend not to help decide what product to offer.

 

VisiRule not only supports the design, documentation and execution of the dense chain of reasoning that is a common feature of insurance underwriting, but it also supports solutions that are scalable from internal users to outside business partners and ultimately to customers.

 

VisiRule’s approach of "programming via flowcharting” permits rapid prototyping and review, and greatly enhances the accurate capture and shared understanding of decision flows. The ability to represent complex decisions in layers helps ensure quality, transparency and ease of maintenance.
 

Explain Your Decision

By tracing the user's session, it is possible to reconstruct the reasoning behind any conclusion reached based on which questions were answered and which rules 'fired'. In this way, VisiRule can produce a justifiable explanation for any decision which is essential to have a transparent and even-handed system. Contrast this to neural nets which can NOT offer any explanation for how a decision was reached and are very much an unaccountable 'black box'.
 

Big Data & Decision Automation

Instead of getting answers using an interactive questionnaire, VisiRule can extract existing data from databases, spreadsheets, sensors etc to answer the questions in a fully automated way.

​

ChatBot Delivery

You can also deliver your VisiRule logic using an Interactive ChatBot where the questions and rules are interleaved with access to a FAQ KB..

​

Decision Visualization & Analytics

VisiRule collects each visitor's journey and these can be aggregated and analysed using the VisiRule Analytics package.

​

VisiRule guides the author in the creation of a chart and generates the appropriate rules. These are made available for immediate execution and testing along with a user questionnaire which contains all the questions needed. When you are happy with your logic, you can upload and publish your chart. The presentation of questions can be customized using standard web technology to reflect corporate branding. As the user answers the questions, the rules are fired and the logic executed. This determines the relevant policy inclusions and exclusions. The chart can be displayed at run-time, like a map, to help users better understand their progress through the complexities of the insurance process.

 

VisiRule enables non-programmers to rapidly build and deliver highly interactive systems which give tailored proposals  of a consistent and assured quality.

 

So Why not use Excel?

Excel has primitive "if-then" capabilities, but are complicated to create, maintain and enhance for decisions of even moderate complexity. ("if state is A and class is B and the sub-class is C and the product is D and the policy effective date is after 7/1/2015 but earlier than 5/15/2016, then do ...").

 

The insurance demo is designed to show how VisiRule can be applied to the very early stages of general insurance underwriting, where there is often little in the way of automated decision support, but where a number of important conclusions can and should be drawn quickly and correctly, even with only limited risk information, e.g.,

        - Availability/unavailability of certain products in certain geographies

        - Warnings that will save the agent and company embarrassment later on (e.g., the company insures this particular class of contractor, but only on the basis that asbestos liability is excluded)

        - Next steps in the process (certain supplemental applications must be completed, and large building values require the company to lay off some of the risk to a re-insurer, and this will normally add a week to the placement process and some additional expense).

        - Advice for the second round of information gathering (e.g., if the insurance applicant is a restaurant, the agent should verify that the restaurant has resolved all prior issues with the local health department).

The business value of automated decision support at this stage of underwriting:
        - Many decisions can be pushed to the least experienced link in the decision chain (prospective customers can get answers without wasting the time of agents, and so with agent-underwriter, junior underwriter - senior underwriter links).
        - Company staff get correct answers quickly without wading through extensive and often poorly organized document repositories and training manuals. (Some 38% of the work of knowledge workers is spent searching for the right information. In addition, correcting mistakes becomes more costly and embarrassing the later in the policy cycle they are uncovered.)
        - Human expertise may not be available for consultation when needed to satisfy the time demands of agents or customers.
        - Junior staff can use decision support systems as training tools to reinforce knowledge that is only applied infrequently, enhancing their sense of mastery and job satisfaction.
        - Claims costs can be reduced if the policy goes out correctly (e.g., with a flood exclusionary endorsement in the event of a weather-related claim).

 

An expert decision is a conclusion or set of recommendations in a routine business context that requires a period of information collection, consultation and application of industry and company-specific process knowledge and rules, as well as “rules of thumb”. Decisions that lend themselves to automation with this class of software tool are part of the daily routine of a professional. Virtually all phases of the insurance delivery chain require expert decisions at key points.

 

If you are involved in the insurance industry, be it marketing, sales, underwriting or claims, you can use VisiRule to automate some of your expert decisions and free some of your best people from some of the more mundane, routine but important tasks, and let them focus on the more complex challenges which are not so easy to automate.

​

LPA, the company behind VisiRule, has been in the business of delivering solid, efficient and reliable rule-based systems since it was founded in 1981. LPA has worked with many partners in the insurance sector on various projects including intelligent underwriting and re-insurance.