Mirela Dimofte

Read the article here: https://iireporter.com/itc-briefing-sass-franklin-manchester-on-governing-ai-responsibly/ Our Take Insurers are racing to use generative AI while neglecting the basics of data integrity and governance. SAS warns that people irrationally trust newer generative tools more than long‑standing machine learning, a gap that has already produced fabricated reports and biased lending outcomes. For an industry built on trust, carelessness is dangerous. The immediate priority should be governance: measure models, monitor drift, document data lineage, and assign clear accountability before granting AI any autonomous role. Manchester urges a practical shift in workflows — move to “AI in the loop” so tools relieve overloaded staff and create time for training, rather than replacing judgement or obscuring responsibility. Strategically, insurers should treat AI as a tool to support existing business objectives, not a separate strategy to chase. SAS’s stance is pragmatic: stop pursuing the next novelty and focus on governing, measuring and using AI in ways that protect customers and the workforce.

Read the article here: https://iireporter.com/itc-briefing-vitesse-ceo-on-modernizing-insurance-financial-infrastructure/ Our Take Vitesse, now authorised as a limited‑purpose trust company by the New York State Department of Financial Services, has launched a regulated payments and funds‑management platform in the US. It virtualises and centralises carriers’ scattered loss funds—balances held with tier‑one banks—so insurers and TPAs can see and control delegated accounts in real time. The platform also automates claims disbursement and reconciliation, cutting costs and settlement time; Vitesse cites clients moving from paper checks to near‑real‑time promises. Its pitch is operational control—carriers can withdraw funds from underperforming counterparties—and it’s expanding a model proven in London and Lloyd’s into US insurance markets while positioning itself as complementary to existing payments players. Those claims are credible but conditional. The limited‑purpose trust charter narrows what Vitesse can offer; integration with legacy TPA and carrier systems will be labour‑intensive, and some intermediaries may resist losing custody and influence. Holding balances at major banks is sensible but not a unique moat, and regulatory and operational risks in the US could slow adoption. For insurers seeking tighter, faster control of delegated funds, the solution looks promising, provided Vitesse executes well and achieves broad industry buy‑in.

Read the article here: https://iireporter.com/itc-briefing-one-inc-extends-its-reach-and-reinvents-claims-payments/ Our Take One Inc is reporting rapid growth (40–50% year-on-year) and is moving into Canada via Guidewire’s Marketplace, planning ClaimsPay availability in 2026. The company cites heavy transaction volumes — about $350m a day — as evidence that its payments network is scaling. It is widening use cases across lienholder, subrogation and mortgagee payments and has a deal with Copart aimed at shortening 5–8 week lienholder delays. The vendor network exceeds one million, and the firm claims integrations with 56 core systems. Those numbers are impressive on paper, but success depends on carriers and core vendors actually deploying and maintaining the integrations at scale. One Inc has combined inbound and outbound flows into a single digital wallet for premiums and claims and added Apple Pay and Google Pay for fast catastrophe payouts. That lowers friction for consumers, yet introduces practical questions around fraud controls, identity verification and real-world uptake. The company is investing in automation, AI and SDKs to speed support and deployments, while noting insurers remain cautious. Its stated mission is to make insurance payouts fast and communicative, but the real test will be sustained carrier adoption, the quality of partner integrations and how it handles regulatory, fraud and operational risks.

Read the article here: https://sloanreview.mit.edu/article/the-three-obstacles-slowing-responsible-ai/ Our Take New York City’s October 2023 AI action plan promised accountability, fairness and transparency and even created an algorithm management and policy officer role. Within months, a city chatbot meant to advise small businesses was found giving misleading, potentially unlawful guidance, exposing weak governance, poor deployment controls and gaps between policy and practice. That episode mirrors a wider problem: companies loudly announce responsible AI frameworks yet struggle to make them real. Systems keep producing biased or opaque results, and regulators from the EU to Canada and South Korea are moving to force clearer standards and oversight. The real risk is institutionalising errors and ethical breaches at scale, not merely technical mistakes. Interviews with more than 20 AI leaders and ethics officers show many frameworks are reputational window dressing. Cultural and structural barriers prevent consistent implementation. The authors highlight three recurring gaps — starting with an accountability gap — and offer practical strategies to close them, signalling that rhetoric must be matched by concrete governance, resourcing and measurable controls.

Read the article here: https://iireporter.com/sapiens-launches-decision-analytics-to-enhance-ai-decision-transparency/ Our Take Sapiens’ new Decision Analytics module adds real-time visibility and monitoring to Sapiens Decision, capturing decision outcomes and the logic behind them, integrating with analytics and BI, and exposing results for natural‑language exploration and scenario testing by business users. The vendor presents it as the closing link in an AI stack that includes ModelAI and IntegrateAI, and as the foundation for a future OptimizeAI capability that will add predictive and prescriptive tuning. The platform is pitched for high‑volume declarative decisioning and is being touted after a Forrester recognition. The capability could materially improve oversight and iteration of automated decisions, but much of the value is conditional. Claims about “capturing every decision” and enabling non‑technical optimisation underplay integration, data quality and governance work that enterprises must still do. Explainability and control matter, yet their usefulness will depend on how transparent the extracted logic actually is, how easily business users can act on insights, and how the future OptimizeAI features perform in practice. This is a sensible step in Sapiens’ AI roadmap, not a plug‑and‑play cure for model risk or operational complexity.

Read the article here: https://iireporter.com/talkdesk-expands-insurance-strategy-with-agentic-ai-platform/ Our Take Talkdesk is shifting from a cloud-first contact-centre vendor to a broader automation partner for insurers, rolling out Talkdesk AI Agents for Insurance to run policy and claims tasks end-to-end: authenticating policyholders, assessing fraud risk, collecting claim details and compiling adjuster-ready files. The platform builds on earlier intent-based automation and generative AI, layers multi-agent decisioning, and ties into core systems such as Guidewire, Duck Creek and Salesforce. Talkdesk points to faster handling, lower abandonment and better employee experience across a mix of legacy and digital-first insurers, and argues that reliable automation will steer more customers to self-service while trimming operational friction. Early integrations with established policy and claims systems are a pragmatic advantage that could speed deployments and protect existing IT investments. Scepticism is warranted. “Agentic” AI claims risk overstating readiness: accuracy on fraud detection, complex underwriting decisions and sensitive customer interactions will need robust human oversight, transparent audit trails and tight data controls. Integration complexity, regulatory scrutiny and change management in emotionally charged contact environments will determine measurable gains. Carriers should treat pilots as experiments with clear KPIs—expect incremental improvement rather than instant transformation.

Read the article here: https://iireporter.com/nationwide-commits-1-5-billion-to-accelerate-ai-and-technology-modernization/ Our Take Nationwide has pledged $1.5 billion through 2028 to speed technology and AI changes, including $100 million a year for the next three years, adding to a $5 billion modernisation drive since 2015. The plan targets claims, pricing and customer service, pushes a 24/7 operating model and promotes human–machine collaboration via tools such as Claims Log Notes, digital twins, Nationwide Notetaker and Copilot Studio. The insurer stresses enterprise-grade security and governance, with separate Blue (innovation) and Red (risk) teams to vet deployments. Executives frame the spend as people-connected and machine-enabled, aiming to free staff for judgement and empathy while automating routine work and sharpening risk modelling. The move is a clear strategic wager rather than mere experimentation, but outcomes will hinge on measurable returns, transparency around data use and bias, and how workforce roles change. Claims of responsible deployment and resilience need independent verification; watch for published KPIs, audit results, customer impact and vendor risk as the programme scales.