3 ways insurance underwriters can gain insight into generative AI | Insurance Blog
Generative AI (GenAI) has the potential to revolutionize the insurance industry by providing underwriters with valuable insights in areas 1) risk controls, 2) structural details and location and 3) insured operations. This technology can help underwriters identify more value in the shipping process and make better quality, more profitable underwriting decisions. Increased rate accuracy from CAT modeling means better, more accurate rates and reduced premium leakage. In this post, we will explore the areas of opportunity, the power of GenAI, and the potential impact of using GenAI in the insurance industry.
1) Risk management details zone in material data
Generative AI allows rthe details of the sk control analysis that should be highlighted to demonstrate the loss prevention measures in place and the effectiveness of those controls in reducing the loss potential. These are critical to informed underwriting decisions and can address frequently missed areas or pain points for underwriters in data collection. Currently when it comes to submission review, underwriters cannot review all submissions due to high volume and different sources. Generative AI allows them to analyze the completeness and quality of all shipments at scale. This means that they go from a limited ability to compare information on similar risks to a situation where they have comparable information on risks by evaluating submissions against the UW Guidelines and current business literature.
What productive AI can do:
- Produce a complete narrative of the entire accident and its alignment with the appetite and book of carriers
- Flagging, finding and identifying critical data required
- Managing updated data lists
- Enrichment from sources that support TPAs/external data (eg, publicly listed products/services of an insurance operatorations)
- Verifying transmission data from those additional sources (eg, geospatial data to verify plant management/proximity to building materials and roofs)
Combining a shipping package with third-party data in this way allows it to be presented in a meaningful, user-friendly way that ultimately aids decision-making. All this can be eenable faster, improved pricing recommendations and risk mitigation. Supplementing information obtained from the broker with third-party data also eliminates the long times caused by going back and forth between underwriters and brokers. This can happen quickly for all shipments at once, prioritized within seconds across the entire portfolio. What an underwriter can do in a week can be done quickly and consistently while making informed, structured recommendations. Underwriter You will quickly know the control gaps based on the submitted data and where there may be major deficiencies / gaps that could impact potential losses and performance costs. That’s right, ttherefore it should be considered in conjunction with the individual insured’s risk appetite. This development ultimately creates the ability to write more risks without higher premiums; to say yes when you could say no.
2) The details of the structure and location help the accuracy of exposure to hazards
Let’s take it example of a multi-building restaurant chain that our insurance company underwrites to show the details of the building. This The restaurant is located in CAT’s favorite location of Tampa, Florida. How can this information be used to supplement the submission to ensure that the underwriter has the full picture to accurately predict the risk exposure associated with this area? The most dangerous hazards in Tampa, according to FEMA’s National Risk Index, are hurricanes, lightning and tornadoes. In this examplei the insurance company had it I used a medium risk level for the restaurant because:
- failure of previous security checks
- lack of storm protection units
- a possible link between previous maintenance failures and the loss event
all of which increased the risk.
On the other hand, to prepare for these risks, the restaurant used several mitigation measures:
- mandatory hurricane training for every job
- metal storm shutters on all windows
- Secure outdoor items such as furniture, signs, and other items that may become projectiles in high winds
All this added to the post shows that they have the necessary response measures to reduce the risk.
While structural details reveal what is actually guaranteed, spatial details reflect the context in which the structure operates. RSk control analysis from building inspections and safety inspection reports unfolds information showing which areas are the driving areas with the most losses, whether past losses were the result of a covered accident or a lack of control, and the adequacy of existing control systems. In the case of a restaurant for example, it it did not have its own storm protection units but according to detailed geo-location data, the building is located approximately 3 miles from the nearest fire station. What this really means is that in terms of content collection, underwriters go from not being able to triangulate from a large volume of information and submissions to being able to drill down to find more context on details within seconds. This allows underwriters to identify and track leaky drivers from data and content aggregation to recommend more effective risk mitigation actions.
3) Performance details to help provide additional risk management recommendations
Details of insured transactions include information from seller submissions, financial statements and information where factors are not included in the Accord forms / seller applications. The dangerous distances of each area associated with the insurance operation and the primary and secondary SIC codes will also be provided. From this, Quick visibility into loss history and driving areas for high losses compared to total exposure will be enabled.
If we take the example of our chain of restaurants again, it can be said that the risk value is ‘high’ instead of the ‘medium’ mentioned above due to the fact. that i The area has potential hazards from eg food delivery operations. Through occupational exposure analysis, this is how we identify that major risk in cooking:
The maximum population density is over 1000 people, and it is located in a shopping area. The number of claims over the past 10 years and the average number of claims can also indicate a higher risk of accidents, property damage, and credit problems. Although some risk controls may be implemented such as OSHA compliant training, security guards, hurricane response training and fire training every 6 months, may be availablefother necessary controls such as specific risk controls for food operations and fire safety measures for an outdoor pizza oven.
This additional information is important in calculating the actual risk exposure and assigning the appropriate level of risk to the customer’s situation.
The benefits of generating AI over underwriting decisions are more profitable
As well as aiding in more profitable underwriting decisions, this information provides added value as they see train new underwriters (with greatly reduced time) to understand data / guidelines and risk information. They improve the accuracy of calculations / accuracy of estimates by pulling all the complete, accurate data for submission to CAT models for each accident and reduce the large discrepancies between actuaries / values / underwriting risk information.
Please see below a summarize the potential impact of Gen AI on underwriting:
In our latest AI for everyone vision, we talk about how productive AI will transform work and reinvent business. These are only 3 ways insurance underwriters can get insights from generating AI. Watch this space to see how generative AI will transform the insurance industry as a whole over the next decade.
If you would like to discuss further, please contact me here.
