Growing Footprints of Technology in the Insurance Sector

概览

This episode of Data Science with Sam discusses how technology has reshaped insurance, especially through digital transformation, cloud infrastructure, APIs, Excel-based workflows, and newer AI tools.

Guest Nick Blamer, Senior Technology Director at Coherent, frames the industry’s technology evolution as a recurring shift toward empowering business users: first by moving mainframe processes to desktops, then by moving workloads to the cloud, and now by turning business logic into reusable APIs.

The conversation also covers technical skills for actuaries, underwriters, and risk managers, including R, Python, SQL, and basic programming concepts. The final part focuses on generative AI, model bias, regulatory guardrails, and career advice for new technology graduates entering insurance.

分段落总结

[00:03] Episode Introduction

[事实] The host introduces the podcast as an episode of Data Science with Sam.

[事实] The stated topic is the growing footprint of technology in the insurance sector.

[事实] The host says digital transformation, big data, insurtech, risk assessment platforms, and infrastructure modernization have changed the insurance industry.

[01:30] Guest Background

[事实] Nick Blamer introduces himself as Senior Technology Director at Coherent, a global technology company primarily based in New York.

[事实] He says he began working in insurance technology in the late 1990s, first developing software as a student.

[事实] His career included consulting companies, large insurance and reinsurance companies, Microsoft, and now Coherent, where he works on software used largely by actuarial users and Excel users.

[03:20] Technology’s Long-Term Shift in Insurance

[事实] Nick describes an early technology shift from mainframe processes that took weeks to desktop-based calculations that could run in hours.

[事实] He says the next major wave moved workloads to the cloud for speed, lower cost, and scalability.

[事实] He identifies the current wave as the broad implementation of APIs, while noting that APIs still require people or tools to build them.

[推测] Nick’s main argument is that insurance technology has repeatedly moved computation closer to business users while making systems faster and more connected.

[05:11] Excel as a Development Platform

[事实] Nick says Excel is the world’s largest development platform and calls it an IDE and first-generation analytics tool.

[事实] He acknowledges Excel has problems, but emphasizes that many people already know how to use it.

[事实] He explains that Coherent Spark can take an Excel spreadsheet, identify inputs and outputs, and generate version-controlled, auditable APIs that can run much faster and scale in the cloud.

[推测] The episode presents Excel not as a legacy tool to discard, but as a familiar interface that can be upgraded into governed, scalable infrastructure.

[07:05] APIs as Reusable Business Logic

[事实] Nick compares APIs to Lego blocks because the same function or calculation can be reused repeatedly and connected with other systems.

[事实] The host says APIs have been revolutionary because they integrate platforms, improve efficiency, and make it possible to build products such as those using OpenAI APIs.

[事实] The host argues that Excel will remain important in insurance even as Power BI, Tableau, and other analytics or visualization tools are used.

[08:48] Connecting Excel Logic to Enterprise Systems

[事实] Nick says Excel and Spark can connect with Microsoft Power tools, Tableau, SAP, mainframes, and other core systems that can use APIs.

[事实] He says business units can own business logic while IT connects the API plumbing and promotes systems to production.

[事实] He gives examples such as valuation, projections, underwriting, and sales as areas where the same calculation can be reused.

[推测] This approach is positioned as a way to reduce back-and-forth between business and IT over requirements and implementation details.

[11:36] Technical Skills Beyond Excel

[事实] The host asks what technical skills actuaries, underwriters, and risk managers should learn beyond Excel.

[事实] Nick recommends comfort with R, Python, SQL, and basic programming concepts.

[事实] He says new programming languages and low-code or no-code tools appear frequently, but many still rely on core programming ideas.

[事实] He says technologies often appear in other industries before later rolling into insurance.

[14:37] Programming Fundamentals for Insurance Professionals

[事实] The host agrees that people should understand fundamental programming concepts even when using low-code or no-code platforms.

[事实] The host says actuaries, risk managers, and underwriters do not need software-engineering-level skills, but should be able to perform analysis or due diligence themselves.

[事实] The host mentions R, Python, and SAS as tools with relevance in insurance, while saying Python covers many broader use cases.

[推测] Both speakers treat technical literacy as a way for business specialists to reduce dependency on IT for routine analytical work.

[16:47] Generative AI and Insurance Governance

[事实] The host asks whether generative AI will have a long-term impact on insurance or whether the industry will return to traditional automation approaches.

[事实] The host raises concerns about data security, privacy, operational risk, regulation, compliance, and AI governance.

[事实] The host says some insurance companies are hiring AI-related legal or compliance roles to prepare governance frameworks.

[18:19] AI Is Already Embedded in Everyday Tools

[事实] Nick says AI is already present in tools such as Microsoft Office, including spell check, grammar correction, and suggested text.

[事实] He says generative AI can help with chatbots, blogs, and basic reporting from collected data.

[事实] He stresses that guardrails are needed so AI use remains legal.

[推测] Nick sees AI less as a sudden future event and more as an existing capability that will become increasingly embedded and less visible.

[19:39] Risk Scoring, Bias, and Legal Constraints

[事实] Nick says AI can be very good at creating risk scores for individuals or objects.

[事实] He warns that not all AI-generated risk assessments are legally permissible, especially when insurance must avoid prohibited factors.

[事实] He gives the example that California P&C insurance rates cannot vary by gender.

[事实] He says AI may reintroduce gender through other data sources, so companies may need to identify and remove those effects from the model.

[22:06] Model Bias and AI Regulation

[事实] The host says model bias is a major issue when models are trained on demographic information such as gender, area, and ethnicity.

[事实] The host says insurers must comply with state laws and avoid biased model prediction results.

[事实] The host argues that AI governance, regulation, and compliance are necessary before taking a major leap in AI use.

[事实] The host mentions examples where ChatGPT can be integrated with Power BI or Excel to help users write formulas.

[24:45] AI Productivity and Misuse Risks

[事实] Nick says AI can help users do tasks they would normally search for online and place results directly into an Excel spreadsheet.

[事实] He says production use still requires controls and understanding of AI risks.

[事实] He says AI can be useful in BI reporting because it can detect patterns and interconnections that may not be obvious.

[事实] He cautions that correlation does not necessarily mean causation.

[推测] The discussion presents AI as useful for productivity and insight discovery, but risky when people accept outputs without statistical or regulatory scrutiny.

[26:45] Avoiding AI Hype

[事实] The host says organizations should step back and ask why they need AI and whether it will reduce operational risk before implementing it.

[事实] The host notes that many existing tools already use AI and that AI is not required to do everything.

[事实] The host returns to Coherent’s Excel API product as an example of technology that can make insurance work easier without requiring every solution to be framed as AI.

[27:52] Career Advice for New Graduates

[事实] Nick advises new technology graduates interested in insurance to explore actuarial exams and notes that technology-related tracks now exist.

[事实] He says the Society of Actuaries has relevant paths on the life side, with similar structures on the P&C side.

[事实] He recommends online courses, communities, chat groups, and learning resources for languages and tools such as low-code platforms, R, Python, and C.

[事实] He says people can explore these areas while keeping their day jobs and use the process to decide what interests them.

[29:40] Networking and Professional Conduct

[事实] Nick says LinkedIn is useful for staying connected.

[事实] He describes insurance as a small world where people often encounter former colleagues again.

[事实] He advises treating everyone with respect because professional paths may cross again later.

[事实] The host recommends listeners look at the Society of Actuaries and says a link is provided in the caption.

[31:22] Closing

[事实] The host thanks Nick for sharing his viewpoints on the questions discussed.

[事实] The host says future podcast episodes will feature academic or industry experts.

播客点评/总结

[推测] The episode is most valuable for insurance professionals who work with Excel, actuarial models, underwriting, risk management, or analytics and want a practical view of how APIs, cloud platforms, and AI affect their daily work.

[推测] Its strongest point is the connection between familiar business tools and enterprise technology: instead of treating innovation as only new platforms, Nick explains how existing Excel logic can become scalable, governed, reusable APIs.

[推测] The main limitation is that the discussion stays high-level and product-oriented in places, especially around Coherent Spark and AI governance, without deeply examining implementation costs, adoption barriers, or concrete case studies.

[推测] The episode is especially suitable for early-career insurance technologists, actuarial students, and business users who want to understand which technical skills matter and why AI should be approached with both interest and caution.