In 2024, AI, GenAI, and ML are stepping into just about any industry you can think of and leaving a trail of exciting transformations behind. Are you actually surprised? Given today’s fast-paced development of tech, we surely are not.

So it shouldn’t come as a shock either that these transformations have reached the insurance arena as well. Specifically, if you’re in underwriting, you’ve likely noticed how these technologies are changing the game, from transforming risk assessments to streamlining processes and enhancing decision-making. But they’ve also brought some challenges with them.

That said, let’s take a closer look at this tech in insurance underwriting.

Advancements in Underwriting

When it comes to Generative AI in underwriting, it dropped some things into this field, and it looks like they’re going to make quite a big splash. Its impactful waves are going to hit several areas. Let’s see what these significant advancements are.

Risk Analysis and Pricing

With AI-powered systems, we’re now able to sift through tons of data—like details about applicants, potential risks, and past records. This means underwriting decisions are not just more accurate but also consistent and unbiased.

GenAI and machine learning models can analyse huge amounts of data, helping insurers assess risk far more accurately, as well as set the best pricing. These technologies make it possible to:

• Break risks into more detailed segments using algorithms like decision trees and random forests
• Automatically check application data against underwriting rules
• Detect subtle risk factors and patterns that human underwriters might miss
• Create dynamic pricing models that adjust based on real-time data.

Operational Streamlining

GenAI has made the underwriting process a lot smoother. How? It did so in several ways:

1. Automated Data Extraction: GenAI takes care of data entry, reducing mistakes and freeing up time.
2. Policy Document Generation: AI has the potential to take over drafting policy documents and contracts, streamlining the process and keeping consistency across all policies.
3. Workflow Optimisation: AI tracks underwriting workflows, pinpoints bottlenecks, and offers up ways to make things run more smoothly.
4. Feedback loops: Companies are collecting more and more data, and the latest AI and ML tech allows them to make better use of it and plug into feedback loops (data-driven or data-informed companies).

Underwriter Copilot

Generative text solutions are delivering chat-like applications for underwriters to offer quick access to detailed answers from underwriting manuals. Capable of understanding intricate questions, these tools gather relevant information and deliver personalised responses. This means underwriters save a ton of time that would otherwise be spent flipping through different manuals and guidelines.

GenAI and chat applications are one thing. And while these are powerful tools for underwriters, there are many additional data sources and insights available to underwriters thanks to growing data sets often processed by various ML algorithms.

Here are a few key ways data and machine learning are augmenting underwriting, moving past just GenAI chatbots:

Alternative Data Sources
Underwriters now have access to a whole new range of non-traditional data, including things like:

• Social media activity
• Geospatial data
• IoT device data
• Public records
• Satellite imagery
• Psychographic profiles
• Smartphone usage patterns
• Market trends

All of these data sources help create a more holistic view of risk.

Automated Data Extraction
Machine learning models can automatically dig through unstructured documents like PDFs, emails, and loss runs to pull out key information. Thanks to this data extraction automation, underwriters can save valuable time on info gathering and input, focusing on more important tasks instead.

Predictive Analytics
Advanced ML models analyse past data to forecast future risks, claims likelihood, and ideal pricing. This leads to more accurate and finely tuned risk assessments thanks to the ability to explore more projections than ever.

Real-Time Risk Assessment
IoT devices (see our article on IoT in damage reduction) and connected sensors make it easy to continuously keep an eye on insured assets. This means insurers can adjust risk pricing dynamically and take steps to prevent potential risks.

Automated Underwriting for Simple Risks
When it comes to simple policies, machine learning models can handle the whole underwriting process on their own, greatly boosting efficiency.

Fraud Detection
AI models can flag potentially fraudulent applications by picking up on odd patterns and irregularities across large datasets.

Customer Behaviour Analysis
A tool that can understand your customers better than ever? That’s what machine learning does. By examining customer interactions and behaviours, ML helps businesses not only predict retention but also set pricing and identify cross-selling opportunities.

Automated Compliance Checks
AI systems act, in a way, as a safety net. How so? They help ensure that underwriting decisions stay in line with both regulatory standards and company guidelines.

Risk Clustering and Segmentation
Machine learning techniques group similar risks together, making it easier to create more tailored underwriting strategies that fit the unique needs of each segment.

Enhanced Decision Support
AI systems may not be fully automated, but they equip underwriters with valuable data-driven insights and suggestions to support their decision-making about complex risks.

---

When underwriters mix different data sources and machine learning tools with generative AI, they get a much clearer and more detailed picture of risk. This not only helps them make underwriting decisions that are spot-on and efficient but also adds a personal touch that benefits everyone involved.

Trust and Transparency in Underwriting Processes

In 2024, trust and transparency have become even bigger priorities in AI-driven underwriting. Insurers are working to address challenges like data inaccuracies and misinformation, often caused by GenAI “hallucinations” (more on this later on). This shift is happening as insurers work to juggle the benefits of AI with the concerns of stakeholders, including underwriters, regulators, and policyholders.

Explainable AI (XAI)

More and more insurers are turning to explainable AI models to make their underwriting processes clearer. These models give straightforward explanations for their outputs, helping both underwriters and customers understand what’s behind each decision. With XAI, insurance companies aim to:

• Provide easy-to-understand models that show key factors contributing to underwriting decisions
• Keep detailed decision records for audits, reviews, and potential due diligence
• Ensure human oversight, allowing underwriters to step in and adjust AI recommendations when needed.

Data Integrity and Quality

Insurers are doing their best to address data inaccuracies and misinformation by putting strong data governance policies in place to ensure quality, privacy, and security. They’re also making it a priority to regularly review and update these policies to keep up with changing regulations. On top of that, insurance companies are paying close attention to external data sources, understanding where the data comes from, how often it’s updated, and how it ties into risk assessment.

Addressing GenAI “Hallucinations” and Misinformation

The growing use of generative AI has spawned new challenges, especially the risk of so-called “AI hallucinations”. What’s that? When AI creates information that seems believable but is actually made up or incorrect, we refer to this as “hallucinations”. These occur when GenAI models produce content that’s not based on real data.

This issue has become a growing concern for insurers, as it can result in misinformation and inaccuracies in the underwriting process. Given the current state of GenAI development, it’s impossible to get rid of hallucinations completely, but the good news is insurers have their ways to keep these risks at bay:

• Being open with clients about the data used for training AI models
• Using strict validation processes to double-check AI-generated information at multiple levels
• Combining AI insights with human expertise to ensure everything is accurate and reliable.
• Introducing manned mandatory quality and compliance checks

Building Trust with Stakeholders—Underwriter Confidence

Even though insurance execs are optimistic about AI’s potential to improve underwriting and reduce fraud, many underwriters aren’t quite sold on the idea yet and remain on the fence about relying on AI-driven tools.

To change this, insurers are getting underwriters involved early in the AI process to build trust and offering training to help them pick up skills in data analysis, machine learning and AI. What’s more, insurers are also making it clear that, when it comes to handling complex cases and ensuring ethical standards, the human touch will always matter.

Regulatory Compliance and Ethical Considerations

Insurance companies are taking a proactive approach by getting ready for new regulations that promote the ethical use of AI in underwriting processes. They’re committed to ensuring their AI systems meet data protection laws and ethical standards. This means following important guidelines like GDPR, CCPA, HIPAA, and the EU’s AI Act.

In addition, insurers are putting strong security measures in place to keep personal and financial information safe and sound. Plus, they’re not just stopping there—they’re also conducting regular audits and vulnerability checks to make sure they’re doing everything possible to protect sensitive data.

Conclusions

With AI, GenAI, and machine learning turning things around in underwriting, it’s clear that insurance is headed into a new era. But it’s not just about doing things faster—it’s about doing them smarter. We’re seeing a shift in how risks are assessed, pricing is done, and decisions are made—bringing more accuracy, efficiency, and personalisation along the way.

There’s no going back now. To stay competitive, insurers need to jump on the new tech bandwagon and embed these advancements into their business. Sure, you could go with a COTS product, but many of them are still catching up. Instead, how about putting some focus on infrastructure, training, and creating a custom approach? Setting up an R&D Team to explore what works best for you could be an option here.

Looking ahead, there’s a challenge lurking on the horizon—keeping the balance between innovation and trust. We need to make sure we don’t lose the human touch while embracing these innovations. The journey is just beginning, and it looks like it’s going to be one heck of a ride—surely a bumpy one at times, so you’d better buckle up and make sure your business is ready to navigate these changes to stay on track.

A partner who’s all about technology might help with that ;). Contact us to tap into our industry expertise to help your business grow.