Looking to Bring AI Into Your Processes? Standardize Your Data First

For a smooth, successful transition to AI and automation, professionals recommend that you consider the quality of your data first.
By: | March 22, 2024

You have to learn to crawl before you can walk and walk before you can run.

The same is true when it comes to modernization efforts in the insurance industry. It’s easy to be tempted by the promising results and shiny new buttons AI offers, but an insurer must lay the groundwork before leaping into AI — they must crawl before running.

A critical, but sometimes overlooked, first step before diving into automation and AI is standardizing data and processes. Laying the foundation by establishing good data practices and regulating how work gets done gives insurers a solid basis from which to grow with more exciting digital projects.

There is little doubt AI and automation hold tremendous promise for the insurance industry — and one cost of entry for insurers wanting to benefit from these innovative technologies is taking time to standardize their data and processes first.

First Comes Data and Process Standardization, Then AI and Automation

Many consider data and process standardization the foundational step before undertaking a modernization project using AI or automation.

Andrew Wynn, cofounder and CEO of Ascend, explained why: “When you hear about AI, often you hear about models they are trained on — millions and billions of data points so connections can be made, insights derived and results provided. All these data points need to be tagged, standardized and easily accessible as new information is provided.

Portrait of Andrew Wynn

Andrew Wynn, cofounder and CEO, Ascend

“Connecting this back to insurance, our businesses often lack that structure from not only a data but a process perspective. For example, different processes for different clients across different team members — this creates an almost impossible data set for any model to know exactly what it is looking at, and that trickles down to AI providing limited to no value.”

Considering the investment of time and money required to implement an AI or automation project, insurers need to get it right the first time and avoid the worst-case scenario Wynn cited, where a costly AI or automation upgrade provides little value to the organization in the end.

Kabir Syed, founder and CEO of Ennabl, provided a road map insurers can use when preparing to implement an AI or automation project.

Advising to “focus on first-party data capture as a priority rather than data that can be easily obtained from other sources,” Syed recommended these steps as best practices:

  • Evaluate data systems to identify improvement areas in the current structure.
  • Set up data governance to ensure quality and compliance.
  • Standardize data models by developing uniform data formats for easy analysis.
  • Use data management tools to automate data cleansing and preparation.
  • Streamline processes by documenting and refining key operations.
  • Enhance data literacy by training staff on the strategic importance of data.
  • Collaborate with technology specialists through partnerships.
  • Design for scalability by building adaptable data and process frameworks.
  • Start small by testing standardization efforts on smaller projects first.
  • Continuously improve by regularly assessing and adjusting standardization practices.

Handling Challenges on the Path to Standardization

As with any substantial project, a data and process standardization upgrade can create challenges for insurers.

Wynn elaborated on some anticipated challenges: “A common challenge we’ve seen is understanding the scope you want to achieve. There are many solutions today that solve specific challenges in the process.

“For example, on the payments side, you may have a vendor to handle billing, a vendor to handle financing, a separate process to handle accounting. Then you may have an entirely separate process to reconcile depending on whether the policy was agency-billed or direct-billed. A key part of automation is consolidation — find partners and processes that let you do more with less. This means you want to broaden the scope of what you’re trying to achieve if you want to really capture the benefits AI has to provide in the future.”

Portrait of Kabir Syed

Kabir Syed, founder and CEO, Ennabl

Listing several other challenges insurers must manage when standardizing data and processes in anticipation of an AI modernization effort, Syed included data privacy and security; integration with legacy systems; maintaining regulatory compliance; and the change management curve of key stakeholders and employees.

Syed provided proactive strategies to deploy to address the challenges insurers face on the path to standardization when embarking on their AI and automation journeys:

  • Implement robust data governance by establishing clear policies and procedures for data management, privacy and security to build trust and ensure regulatory compliance.
  • Leverage hybrid cloud solutions that can accommodate legacy systems while providing the flexibility to adopt new technologies.
  • Invest in data management and cleansing by improving data quality through advanced data management solutions and cleansing processes to ensure AI models are trained on accurate and comprehensive data sets.
  • Foster a culture of continuous learning that encourages upskilling and reskilling among employees to bridge skill gaps and reduce resistance to new technologies by highlighting their benefits.
  • Engage with regulators to maintain open lines of communication to ensure compliance and stay informed about evolving regulations.
  • Prioritize customer-centric AI applications by developing AI solutions that enhance customer service and personalize the customer experience while maintaining a balance with human interaction.
  • Start small and scale gradually by starting with pilot projects to demonstrate value and gain experience before scaling up AI and automation initiatives across the organization.
  • Monitor and adjust through regular reviews of the performance of AI and automation systems, making necessary adjustments to align with business objectives and customer expectations.

Syed and Wynn agreed that data standardization is a critical step in the race toward AI and automation adoption.

“Standardization of process, data is imperative if we want to see the benefits of technology, be it machine learning or AI,” said Syed. “Otherwise, we will not have any ROI.”

Investing in data and process standardization is the prerequisite for a successful AI and automation project, allowing insurers to be ready for the significant changes to workflow the digital transformation to AI and automation will bring. &

Abi Potter Clough, MBA, CPCU, is a keynote speaker, author and business consultant focused on Insurtech, leadership and strategy. She has over 15 years of experience at a Fortune 500 company with expertise in P&C claims operational leadership, lean management consulting, digital communications and Insurtech. As the past chair of the International Insurance Interest Group of the CPCU Society, Abi remains involved in many international initiatives and projects. She has published two books about change management and relocation. Abi can be reached at [email protected].

More from Risk & Insurance