5 Steps Insurance Must Take to Integrate Artificial Intelligence into Its Post-Pandemic Risk Strategy

By: | March 2, 2021

Saravana Kumari Sundaram heads the Insurance Solutions at Virtusa. She has three decades of global experience spanning across Insurance, Securities and Capital Markets and Banking. She started as a technologist and broadened the horizons to large programs and account management followed by an in-depth focus into the insurance industry. She is an Associate from Insurance Institute of India (AIII) in both Life Insurance and General Insurance and a member of Life Office Management Association (LOMA), USA and Chartered Insurance Institute (CII), UK. She is passionate about applying technologies for business disruption. She is an experienced PMP professional and a Six Sigma green belt.

As with most industries, the insurance world was rocked by the arrival of COVID-19, which thrust most businesses into a precarious position — finding new ways to assess vulnerabilities and up the customer experience while operating in a remote world.

While COVID arrived like a chilling blast of arctic air, it didn’t freeze businesses and prevent them from taking action.

Quite the opposite; the pandemic invigorated the need to embrace new digital initiatives including artificial intelligence (AI).

According to a December 2020 study from TransUnion, “digitization accelerated in response to the COVID-19 pandemic as insurers looked to better serve customers and remain competitive in today’s evolving marketplace — digital adoption in the insurance industry grew 20% globally in the past year.”

Now, just weeks into 2021, this focus on innovations such as AI remains a top-tier priority and the key to success in the era of “new normal.”

Here’s why — insurers have always had access to mountains of data but are now being asked to process it at unprecedented speeds that most simply cannot meet.

Enter AI.

AI can take data from traditional and new resources such as Internet of Things (IoT) devices, social media pages and credit reports to provide insights on levels they’ve never seen before.

For example, we’ve seen an email triage system using natural language processing and AI/ML deliver 80% accuracy and an 89% confidence level while reducing turnaround times by 50%. In addition, the system played a pivotal role in fueling better customer experiences and freeing agents to focus on more pressing customer-facing needs.

It’s benefits like these that have given the industry hope, but they don’t shed light on where this AI journey begins. Here are steps that each company must consider as they kickoff their AI initiative.

1) AI Strategy and Roadmap

Gartner predicts that nearly 80% of AI and machine learning initiatives will not deliver business outcomes through 2022.

There are many reasons why this is the case.

One is that businesses fail in their due diligence, specifically when it comes to the following: discovering, analyzing and prioritizing the use cases that address their challenges and then determining, which can be addressed in both the short and long term.

We’re all eager to launch our AI initiative, but we cannot ignore these core steps.

2) Forge Your Data Strategy

Another reason AI investments often miss the mark is they lack an underlying data strategy. Some of the typical risks with AI implementations include:

  • Inadequate understanding of the organization’s data
  • Insufficient availability of the right data
  • Poor data quality
  • Lack of clear ownership

The right data strategy includes many elements such as:

  • Data Diversity: Bring in data from your systems of record as well as alternate and third-party data sources and both structured and unstructured data.
  • Data Glossary and Lineage: Create a glossary where technical and business metadata are captured and continuously updated as new sources are added. This ensures faster identification of the right datasets.
  • Data Quality: Data quality checks address completeness, uniqueness, timeliness, accuracy, consistency and validity of datasets before they are used for AI/ML use cases.
  • Responsive Data Architecture: Cloud-enabled architecture helps to scale and optimize compute capacity usage across on-premise, public cloud or hybrid workloads.
  • Data APIs: Exposing data via APIs simplifies and provides uniform access to data consumers, reduces redundancies, enables agility and self-service, measures consumption and monetizes curated data and insights.

3) Check Your Readiness

Next comes the assessment of current technologies, business and data landscape. This process will allow you to assess your AI maturity level.

Deploying AI requires a massive amount of readiness from a technology and data standpoint. Determine data, model training and deployment gaps before you begin and find ways for how these can be bridged.

4) Train Your Team

Not only do you need to check your readiness from a technology standpoint, you also must get your team up-to-speed on these new AI capabilities.

This means training employees, assessing their skill sets and “up skilling” as needed.

5) AI Governance

Those who are new to AI may be so focused on the benefits that they overlook governance and ethics. Make sure you have a plan in place that ensures the ethical use of data and keeps the team one step ahead of new regulations.

The insurance industry is ringing the AI bell, and the race is on. However, while the urge to finish first is enticing, taking the right steps is what will allow these businesses to do it right the first time and operate on a level demanded by this “new era.” &

More from Risk & Insurance