Case Studies

Design Thinking Helps A US-Based Insurance Major to Look Beyond

Business Need
The client, an American diversified global insurer and the third-largest property and casualty insurer in the United States, wanted underwriting accuracy, improve efficiency with Portfolio optimization & ensure the right coverage for the right customer, Portfolio expansion and ensure standardization of operations & decisioning.

The team was able to extract the unstated requirements as: Operationalization of Intellect’s products within the client’s organization and adapting to varied maturity levels of stakeholders in Artificial Intelligence (AI) implementations.


The Role of Design Thinking
Understanding of problem statement was done by asking ‘Why’ a number of time during workshops. Customer engagement began with identifying and prioritizing the 1000gm items based on complexity and criticality. This proven capability was then extended for other business areas.

The team also created targeted pilot and training sessions for various user personas. ML feedback loops built into the product (Magic Submission product), enabled us to detect user behaviors to mitigate and ensure continuous learning for edge cases.


Listening-Dialogue-Observation was Key
We approached the client’s need by Listening to their strategic direction and pain points. Our team also had a dialogue with the right stakeholders such as senior leaders of Underwriting Strategy, Product Owners, Underwriters, Data Scientists, Process Owners and Operations teams.

Workshops were also conducted to understand the client’s process, roadblocks, pain points, current partners, and articulate/map where and how existing products and technology fits for addressing current and strategic need. In addition we followed the principle of Observation by shadowing of current process of the client’s operations team to gain insights on procedures and more importantly nuances (variances) in human judgment calls (decisions made) for the exact same scenarios.

This resulted in ensuring a ‘start small, scale fast’ strategy for a phased release for varied lines of business and use cases which ensured success and speed of adoption with continuous user feedback loops.


Blind Spots
Employees opined their jobs would be obsolete due to the implementation of Magic Submission. Cultural adaptation of cognitive transformation was initially a blind spot.

Changing the underwriter’s behavior to utilize the risk insights provided by Risk Analyst was a challenge. Embedding the account insights reports into their existing dashboard through API mitigated this.

The stakeholders could not come up with a single end state for certain scenarios. Handholding and challenging them on their expected end state visualization with valid data points helped mitigate this. Measurement and quantification of success criteria for AI & ML projects was new to the client.


A Solution Powered by Design Thinking

Multiple consultative engagements with various stakeholders in a divergence mode were done, to map out the business impact and ensure a minimally disruptive path to successful adoption of tools.

We, then, established data credibility with Liberty’s stakeholders through pilot runs. The tool (RA) found 38% of nearly 12,000 new business prospects were ‘high interest’ accounts that should have focused sales efforts. The tool (RA) also indicated that more than 10% of excluded accounts could be within the client’s risk tolerance.

For Magic Submission, introducing a ‘confidence score’ at the field level, ensured the number of fields business users were required to review was minimal (applied the law of Less is more) thus showcasing high accuracy of the product ~ 98%. For RA we introduced small data packages targeted to specific use cases and easily consumable in extendable data sets as needed.