iMSA Reduces Production Bugs with SonarQube Server

From ServiceNow SKO 2026 so far (more to follow). Agentic AI does not remove the need for new foundations in insurance. It intensifies it. Agents still require structured data, accessible data and the ability to act on this as close to real-time as needed: 1️⃣ They need explicit permissions, and management and to be able to adapt these. 2️⃣ They need audit trails, analysis and reporting. 3️⃣ They need enforcement points, they need controls - security & risk management matters more than ever. 4️⃣ They need contracts between systems that define what is allowed, what is logged, and who is accountable when something goes wrong. And things can go wrong unless responsibility is taken to make things deterministic - intelligence isn't without fallibility. 5️⃣ We need to address the messy middle of insurance, operational transformation is hard but this is where the value will intensify dramatically. Working with the ServiceNow ecosystem in recent weeks has unpacked solutions I believe insurance deeply needs. Focused on outcomes we need to use these toolkits to unpack new value in insurance. Rewiring everything, but 💥 evolving 💥 from today to this tomorrow isn't a constraint. In fact, it will likely yield better future state outcomes. Rapid learning, change as a constant and taking people along this journey. I've been unpacking the need to take what is essential to insurance and how it operates and concentrate on what makes it better. We do need to lay new foundations, but unlike a building or physical infrastructure we can actually restructure entire components as we continue to embrace machine intelligence. It's the vision for a far more adaptive future that matters - your organizations Adaptability Quota (AQ) will define your competitive future. And I can't wait to continue to find out more whilst here and continue to have these discussions. Mike Daly Nigel Walsh Sukhi Gill Bob Moore Rob Gilley Mike Luntz Chirag Jindal Daniel Mooney Christine Olmsted Synechron Calitii RapDev Waivgen Lisa Wardlaw Robert Pick

Policy Fighter — An Adversarial Tool for Insurance Denials Headline: Just open-sourced Policy-Phi Fighter — a practical tool to challenge unfair insurance denials using structured rebuttals Post Body: Insurance companies often rely on automated, high-volume denial processes to wear down claimants. It works because fighting back takes time, expertise, and energy most people don't have. I built something to level the playing field. Policy-Phi Fighter is an open-source prompt framework + simple script that helps users generate precise, evidence-based rebuttals to insurance denials. How it works: Legal Audit — Identifies inconsistencies between the denial reason and policy wording or law (e.g., Insurance Act 2015 in UK). Structured Rebuttal — Produces clear, professional responses citing specific clauses and escalation paths (FOS, bad faith liability). User Guidance — Suggests supporting evidence to strengthen the case. It's not magic — it's systematic application of the insurer's own rules and obligations against questionable denials. UK-specific leverage: Section 13A (reasonable payment timeframe) and implied duty of good faith — powerful when highlighted correctly. GitHub repo live: https://lnkd.in/gQNhDWu3 Includes: Ready-to-use system prompt for LLMs Example denial → rebuttal flows Basic PDF text extractor (JS) Contribution guide for other jurisdictions Public domain — fork it, improve it, use it. Goal: Reduce the asymmetry in consumer vs. insurer disputes. If you've ever fought an insurance claim, you know how exhausting it can be. This is a small step toward making it less one-sided. Feedback welcome — especially from legal pros or anyone who's added their country's equivalent "hammer" clauses. #Insurance #ConsumerRights #OpenSource #LegalTech #Automation https://lnkd.in/gzx84RYF

💼 Insurance Industry Modernization: From RPG Legacy to Java Excellence The insurance industry relies on complex business rules embedded in legacy systems. This global insurance company proved these systems can be successfully modernized without losing critical functionality. Legacy complexity doesn't have to mean modernization complexity. Discover their transformation approach:

After 10 years of building in silence — we're ready to talk "We've spent the last decade building insurance infrastructure. Now we're going to tell you how it actually works." For almost 10 years, we've been doing one thing quietly but consistently: Building CLP Hub. Shipping projects. Helping insurers, banks, and partners move from legacy chaos to cloud clarity. Most of that time, we were too busy shipping to talk about it. We launched products. Fixed edge cases at 2 a.m. Survived regulator audits with clients. Helped teams migrate from Excel spreadsheets and aging PAS systems to modern cloud infrastructure. Now, we're ready to be a bit more open. Why now? Because we've learned that transparency builds trust — and the insurance tech world has enough vague promises and polished decks. What it needs more of: real stories, real timelines, real lessons. What we'll share here: 🔹 How insurers and banks actually go live in 3–6 months — not the sales pitch version, the real one 🔹 What happens between "let's do a pilot" and "we're in production" — the messy middle that no one talks about 🔹 Lessons from projects across Georgia, MENA, and Eastern Europe — what worked, what didn't, and why 🔹 How we think about product, pricing, and partnerships — the trade-offs, the tough calls, the "we'll figure it out together" moments 🔹 What went wrong on some projects — and how we fixed it (because pretending everything is perfect helps no one) Our goal is simple: Be useful for people who are responsible for core systems, digital channels, and embedded insurance. Not to entertain your feed with generic "insurtech thought leadership." If this sounds relevant to you: ✅ Follow CLP Hub to get updates ✅ Leave a comment with topics you'd like us to cover — core migration? API integrations? Pricing models? Pilot-to-production workflows? Something else? We'll listen. And we'll share what we know. Let's make the next 10 years of insurance tech a bit more transparent than the last. Welcome. 👋

Legacy insurance software isn’t just old — it’s expensive. Here’s a look at how outdated systems are consuming IT spend and what needs to change. A must-read for anyone thinking about modernization and cost control. 👇 https://lnkd.in/giEEMPkK Combined Ratio Solutions Michael Jones Luke Magnan Ben Steele Paul Shiman Scott Lea

Chapter Five - part1 Having covered workflow segmentation and prompt engineering in previous chapters, we now turn to the practical handling of individual LLM tasks. To achieve consistent, high-quality results in insurance applications, three key aspects must be mastered: 1. Model Selection Insurance tasks demand deep reasoning and subtle domain nuance. Unless the task is extremely simple, choose models with at least 30B parameters to meet the requirements of complex underwriting or claims scenarios; however, keep in mind that larger models consume more time to generate output. You must find the "sweet spot" between reasoning depth and latency for your specific use case. Furthermore, as discussed previously, we use JSON to manage the data structures passed between models and code, so models with native support for JSON output are preferred. That said, prioritize the model's core reasoning ability; if a powerful model lacks native JSON support, you can simulate the format via text instructions to bypass this limitation. I personally do NOT recommend pre-training a proprietary model from scratch or even conducting heavy Full Fine-tuning/LoRA. The pace of evolution from giants like OpenAI (GPT), Google (Gemini), X (Grok), or Alibaba (Qwen) is so rapid that a specialized model built today at immense cost could easily be outperformed by their next general-purpose iteration tomorrow. More importantly, in the insurance domain, if you have the resources to prepare the high-quality, structured data required for effective fine-tuning, that effort is far better spent building a traditional, high-quality structured knowledge base. Relying on a well-structured database for direct retrieval is significantly more reliable and cost-effective than trying to "bake" that knowledge into a model's weights via fine-tuning. For insurance-specific adaptations, focusing on system architecture and prompt engineering remains the most sustainable and effective path. 2. Parameter Settings Insurance is a rigorous industry that requires stability and reproducibility. In terms of parameter control, I advocate for a strategy of "intentional downgrading" to achieve "deep determinism": >>> l Prioritizing Control Over Built-in Features: Most modern models now feature native Chain-of-Thought (CoT), web search, and tool-calling capabilities, which I personally find impressive for general-purpose use. However, unlike a general programmer who might rely on these features to let the LLM "figure out" the logic, an insurance expert has already embedded precise domain experience into structured logical steps. Consequently, we should disable these features to prevent the model from engaging in redundant, unguided reasoning that introduces uncontrolled variables and reduces the stability of the final judgment. (Continued in next post – Part 2) #Insurance #InsurTech #AI #Claims #Underwriting #LossAdjusting

𝗜𝗻𝘀𝘂𝗿𝗮𝗻𝗰𝗲 𝘁𝗲𝗮𝗺𝘀 𝗵𝗮𝘃𝗲 𝗮𝘂𝘁𝗼𝗺𝗮𝘁𝗲𝗱 𝗮 𝗹𝗼𝘁 — 𝘆𝗲𝘁 𝗺𝗮𝗻𝘆 𝘀𝘁𝗶𝗹𝗹 𝗳𝗲𝗲𝗹 𝘂𝗻𝘀𝗮𝗳𝗲 𝗮𝗯𝗼𝘂𝘁 𝘁𝗵𝗲 𝗼𝘂𝘁𝗰𝗼𝗺𝗲𝘀. That’s because traditional test automation wasn’t built for insurance, where software isn’t just about screens — it’s about decisions, logic, regulation, and trust. In our latest article, we unpack: 🔹 Why UI-first automation can fail quietly in insurance 🔹 How complexity lives in rules, not screens 🔹 What it takes to test correct decisions, not just flows 🔹 How domain-aware automation changes the game 👉 Read the full article and see why modern insurance test automation demands a different approach: https://lnkd.in/g4SGYxFi 💡 Visit our website for more insights on insurance tech, QA strategy, and innovation. #InsuranceTech #TestAutomation #QualityAssurance #InsurTech #AItesting

A $630M insurance startup just gave their entire team Speechify. Corgi is rebuilding insurance from scratch. Direct. Transparent. AI-native. Their team moves fast. The reading load doesn't. Policies. Reports. Industry research. Dense documentation that takes hours to get through. They needed a way to move faster. That's where we come in. Now their team consumes research at 3x speed. Reviews policies in a fraction of the time. Uses voice AI to work through information without being glued to a screen. This is what we built Speechify for. Not just individuals. Teams. Companies. Anyone drowning in text. Congrats to Corgi on the Series A. Excited to help you build the future of insurance. - written with Speechify voice typing

📣 Fulcrum has raised a $25M Series A to rebuild operations infrastructure for insurance brokerages! SPC Founder Fellows Arjun Mangla and Sambhav Anand started with a thesis: insurance makes audacious things possible. Catastrophic loss becomes manageable risk. Yet, brokerage operations ran on fragmented systems. Account managers spent their days toggling between platforms, manually syncing data, chasing down policy details. When they first starting building there were quite a few no's. Until the first yes. Once firms saw how AI could handle the repetitive coordination they brought Fulcrum deeper into their workflows. The team worked to build AI agents that could navigate insurance's conditional complexity: the layered policy terms, the coverage nuances, working with brokerages burned by previous software investments to show them something different. Now Fulcrum runs operations for most of the top 50 US insurance brokers, automating work that previously required continuous manual oversight. The $25M Series A was led by CRV, with participation from SPC, Foundation Capital, and various angels. Congratulations Arjun Mangla, Sambhav Anand, and the Fulcrum team!

Getting the Most from your Managed Service Provider The small to mid-sized insurance agency lives in a complicated and competitive environment. Your clients are constantly under threat of encroachment by competing carriers seeking to build their own businesses by stealing your market share. The last thing you want to worry about is a technology infrastructure failure. The really bad news here is that agency budgets require doing more with less. This rules out hiring a technology team to streamline and optimize your hardware and software. Fortunately, there are IT managed service providers that not only understand your business but can help you grow it. This post explores how to best work with a managed service provider and how insurance firms can establish productive partnerships with these IT professionals. How a Managed Service Provider Can Help