What are the sources of data used by kasko2go’s products?
Our data collection process involves utilizing a variety of sources, which include:
- Governmental data sources: We gather demographic, geographic, and criminal data from official government databases.
- Open sources: We access publicly available information, such as traffic data, road structure details, and location data.
- Commercial sources: In addition to the above sources, we also acquire data from commercial providers, albeit for a fee.
We assure you that our usage of this data is legitimate, and we are dedicated to upholding access to and availability of this information.
How do you customize your calculation models based on client needs and priorities?
We tailor our calculation models based on the key performance indicators (KPIs) that are of utmost importance to our clients. By understanding their priorities, we can customize our approach. Here are a few examples:
- Loss Ratio: We aim to achieve a significant reduction of 4% in the loss ratio, ensuring improved efficiency and profitability.
- Claim Frequency: Our focus is on reducing claim frequency by 9%, which leads to a smoother claims process and lower costs for our clients.
- Additionally, we address other objectives such as increasing market share and optimizing the marketing budget to enhance overall business performance.
How do you collaborate with actuaries and underwriters to identify and enhance specific key performance indicators (KPIs) for clients?
We collaborate closely with the actuaries and underwriters of our clients to determine the specific key performance indicators (KPIs) they aim to enhance. Once these KPIs are identified, we develop a customized function that converts the risk measurements into price coefficients.
After obtaining the analysis results, we offer several options for our clients to apply them in their work. The choice is ultimately made by the client, and accordingly, we adapt and modify the presentation of our analysis results to align with their preferences and requirements.
How do you facilitate the process of obtaining regulatory approval for model modifications?
In numerous countries, making modifications to models necessitates obtaining regulatory approval. To facilitate this process, we offer a comprehensive report that offers a transparent and systematic explanation of our calculations. This report plays a vital role in obtaining regulatory approval as it provides regulators with the necessary information and insights they require.
How is model retraining initiated and pricing changes implemented in the context of insurance?
Model retraining is not performed automatically; it is initiated upon request from the insurance company. Any changes to pricing are implemented only if specifically requested by the insurance company for model retraining purposes.
How long it takes to reduce the loss ratio?
From day one, the insurance company integrates API calls into their operations and incorporates the recommendations derived from these calls into their decision-making processes.
What are the advantages or benefits of the first stage?
We excel in developing highly accurate models and harnessing extensive datasets surpassing those utilized by any other insurance company. As a result, the following outcomes are achieved:
We have a comprehensive manual that includes all the graphics displayed on the dashboard. This manual serves as a reference guide, providing detailed information and explanations for each graphical component.
- Enhanced precision in risk calculation and identification of improvement opportunities.
- Potential for substantial growth for the insurance company.
- More accurate geographical risk distribution, providing valuable insights for risk management.
The initial phase involves receiving and processing the data, serving as the preparatory stage for Phase II. During this phase, the model is not built from scratch but can be customized based on the key performance indicators (KPIs) set by the insurance company, our customer.
What is the process for conducting a retro test, and how do you recommend the insurance company validate and verify the findings?
For our retro test, we utilize data that is provided to us by the insurance company. In cases where the insurance company suspects potential anomalies, we typically recommend that they share 70% of the data with us while retaining 30% for their own use. This reserved portion of data can be used to independently validate and corroborate our findings.
How do you assist the insurance company in connecting to and integrating your API into their IT system?
We offer the insurance company access to our API (Application Programming Interface) along with the necessary API keys. Additionally, we provide comprehensive manuals that guide them through the process of connecting to the API. Our dedicated tech support team is readily available to assist with any issues related to integrating our API into the insurance company's IT system.
What specific specialists do you require from the client’s side for the initial step and for the subsequent step?
For the first step, we would need a specialist who can provide the necessary data to us. Typically, this role is fulfilled by a member of the actuary or underwriter team.
Regarding the second step, we require a member of the IT team from the insurance company who can handle the finalization of API integration (which usually takes approximately 1-3 full-time days). Additionally, we would need one underwriter who can assist in setting up the pricing schema.
What differentiates your work from that of my actuaries?
Our unique approach involves utilizing AI in a manner that ensures transparency, comprehensiveness, and regulatory acceptance of the results.
What are the implications if the address is incorrect or the client does not drive around their home?
In the initial stage of our product deployment, we closely monitor the correlation between the provided address and associated risk. If we observe a systematic issue across the portfolio, such as incorrect addresses or a mismatch between the address and the actual driving location, we will identify this with a low correlation. In such cases, we will promptly notify the insurance company before proceeding with the deployment to production to ensure the accuracy of the calculation.
What sets me apart if your solution is marketed to every company in the country?
The pricing will vary as it is influenced by the biases that have been established by the insurance company's operations to date. Each company has its own distinct customer selection process, considering both risk and marketing perspectives. This uniqueness in customer selection shapes the training data we utilize and forms the basis for customer segmentation.
How does your system respond to significant market changes such as inflation or the impact of Covid?
In the event of significant market changes, such as inflation or the effects of Covid, our system closely monitors various global indicators. For instance, we observe changes in Average Relative Loss, which subsequently affects the average Loss Ratio. By continuously analyzing these indicators, our system adapts and adjusts to the evolving market conditions, ensuring the accuracy and relevance of our calculations and risk assessments.
How do you keep the insurance company's data secure?
We prioritize the security of the insurance company's data and have implemented measures to ensure its protection. We align with the requirements outlined by the General Data Protection Regulation (GDPR) to safeguard personal and sensitive information.
Additionally, our Chief Information Security Officer (CISO) is actively working towards obtaining the ISO 27001 certificate. This internationally recognized certification sets rigorous standards for information security management systems, further enhancing the security protocols and practices we have in place. By adhering to these standards and continuously improving our security measures, we strive to maintain the confidentiality, integrity, and availability of the insurance company's data.
Will working with you replace my actuarial department?
Working with us will not replace your actuarial department. Instead, our AI-based product is designed to support and complement the work of your actuaries. By leveraging advanced technologies and data-driven insights, our product can enhance the capabilities of your actuarial team, providing them with additional tools and resources to improve their effectiveness and efficiency. Our goal is to work in collaboration with your actuarial department, leveraging the strengths of both human expertise and AI capabilities to achieve better outcomes for your organization.
Do we receive analytics that predict the outcomes after a model update?
Following the model retraining, the insurance company receives a report that compares the old and new versions of the model. Based on this report, the insurance company makes decisions regarding the model update. No changes are implemented without the approval of the insurance company.
How can I monitor the results of API calls?
We have full confidence in the accuracy of our calculations and can provide justifications for them. As part of risk control, the insurance company has the option to conduct A/B/X testing to monitor and evaluate the outcomes of API calls. This allows for continuous monitoring and assessment of the results, ensuring transparency and accountability in the process.
Will your recommendations alter my target audience?
Our recommendations are derived from our calculations and serve as a valuable tool for the insurance company to make informed decisions regarding both the existing customer base and potential expansion into new niches. However, it is important to note that the insurance company retains full autonomy and control over the target audience, as well as the marketing and advertising budget. Our recommendations provide insights and guidance, allowing the insurance company to strategically align their efforts while maintaining decision-making authority.
What if the insurance company already has or outsources a Data Science team?
It is common for Data Science teams to have their hands full with their own projects and priorities. By partnering with us, the insurance company gains the advantage of an additional perspective, reducing the time gap between research and production at a competitive price.
Our approach has been extensively tested not only on your insurance company's data but also on diverse datasets from numerous sources worldwide. Furthermore, it has undergone validation by Munich Re AI insurance vehicle, adding an extra layer of confidence in the effectiveness and reliability of our methods.