SortSpoke Blog » Latest Articles
blog featured image
Aug 22, 2023 SortSpoke

How Should We Use AI for Insurance Underwriting?

Decide whether AI Automation or AI Augmentation is right for your team

Over the last few years, there’s been a lot of hype about AI technology revolutionizing insurance underwriting. But for some teams that have tried these tools, that has all it’s been: hype. Often, underwriters end up dissatisfied due to the mismatch between the lofty promises of AI developers and how the software actually works in their day-to-day workflows.

The question insurers must ask themselves is, “How should we use AI?” Do we want or need AI to completely automate workflows? Or should we be leveraging AI to augment our human expertise?

As with all important but complex topics, the answer is “it depends.” This article guides insurance professionals through the AI concepts they should understand to make informed decisions about integrating AI into their underwriting processes.

The bottom line: it depends on how structured your data is

The AI options in insurance and all other markets are evolving fast. Some AI analysts’ predictions of a fully automated future from just a few years ago are already falling short.

Choosing between AI automation (where AI systems entirely manage processes) and AI augmentation (where AI supports and amplifies human expertise) requires understanding each approach's benefits and limitations. P&C underwriters will have different use cases than life insurance underwriters. Those insurers serving enterprises will have different needs than those serving SMBs or consumers.

If your processes and the documents and data you receive are highly standardized, then fully automated AI can make sense. If your workflows are looser or you’re dealing with more unstructured data that could come in hundreds of different formats or free form, then augmentation is probably the more valuable approach.

Download our free self-assessment worksheet to see if your company could benefit from AI-powered document processing.

 

AI automation

The World Economic Forum conducted a thorough study on the jobs and industries most likely affected by AI tools that either automate or augment human-performed tasks. Not surprisingly, they found that routine and repetitive tasks are the ones AI will most likely fully take over from human labor. For example, that WEF study found that over 80 percent of the tasks performed by a credit analyst or clerk could be fully automated. AI could automate 68 percent of a telemarketer’s job.

In insurance, one of the first places business leaders have looked to automate fully is data processing and extraction in underwriting submissions. Automation is possible, but it depends on several factors, including:

  • How standardized the documents they receive are.
  • Your teams need true 100 percent data accuracy from automation.

For the AI to understand what it ‘sees’ in each document, developers and users need to build templates the AI can reliably follow every single time. For example, if your underwriting team only receives four or five types of highly structured submitted forms, then an AI designed for automation will work well.

AI augmentation

“The tasks with the highest potential for augmentation are those that require more abstract reasoning skills, especially those that combine interaction with people,” the WEF report states. In many other instances, AI cannot take over entire tasks, especially involving critical thinking, deep expertise, and complex problem-solving. But they can provide specific, contextual support. That is the goal of AI augmentation: to save underwriters and other insurance personnel time by accelerating overall workflows and peeling off portions of tasks that can be automated.

An AI may not be able to take over tasks involving critical thinking and complex problem-solving, but it will be able to lend a hand. The WEF report says, “This LLM assistance will save workers time and consequently boost their productivity, especially if they have roles involving mathematical and scientific analysis. Insurance underwriters are most likely to benefit from this – 100% of their tasks have the potential to be augmented.”

Large language models (LLMs)—one of the core innovations behind the latest wave of AI tools, like ChatGPT—can deliver a great deal of value to insurance teams, but only in certain use cases. Only a human expert has the deep contextual understanding necessary for complex, analytical decisions. 

One of the issues in commercial lines insurance is that documentation is complex and constantly changing, making data extraction challenging and virtually impossible to automate without any human intervention.

Instead, AI augmentation can increase your current underwriting capacity. With this approach, your underwriters stay involved in the process. Arming your experts with tools that help them process documents faster allows your team to review submissions more efficiently, without sacrificing accuracy and ensuring you're keeping revenue on the table.

By keeping your underwriters in the driver's seat rather than automating the process entirely, you won't have to worry about data quality or increasing the risk of poor writing decisions that might result from full automation. 

Artificial intelligence and machine learning technologies can free underwriters and claim adjusters from manual and repetitive tasks. This technology can empower underwriting and claims teams to focus on higher-value activities rather than replace them. 

Solutions like SortSpoke can help your underwriting team with these tasks. Augmenting underwriting processes with AI like SortSpoke can lead to 40 percent increases in written premiums. It also makes other it easier for underwriters to:  

Investing in your greatest asset—your people

Research from the Brookings Institute showed that properly applied AI automation and augmentation can create a “rising tide” effect, expanding the need for skilled human labor rather than replacing it.

Your employees are your greatest asset, and underwriters have expertise that is hard to replace. With insurers struggling to attract and retain talent, you can leverage AI to empower your team and give them control over how they work instead of trying to replace them or fill gaps by automating the underwriting process. 

How SortSpoke helps

  • Create audit trails and capture higher-quality data for compliance reporting
  • Recruit and retain staff by eliminating manual and redundant processes
  • No technical resources or templates required 
  • Increase document processing speed by 4x-10x 

Book a risk-free, no obligation discovery call with our team to learn more.