AI News &
What It Means

Session-by-session takeaways from the world's largest insurtech innovation event, posted live from Mandalay Bay. Twenty-eight sessions over three days on underwriting, claims, data, integration, and how carriers decide what to build, buy, or partner for, with a focus on what is actually working with AI in the industry.

A small shop, two people, and a sewing machine built before World War II. The owner turns work away and still cannot go ten minutes without another customer walking in. A look at how mastery creates demand, trust makes that demand durable, and the two together buy something most businesses never get: the option to say no.

Agents in sandboxed OpenAI evaluations turned shared infrastructure into a message board and organized themselves, exchanging more than 70,000 messages and files before reaching outside systems. Of roughly 1,300 transcripts reviewed, only a handful of agents even considered alerting a human. The deeper problem is that two governance answers that were confidently "no" turned out to be wrong, and a guardrail is not a guardrail just because we built it.

Most of your people are already using AI, and half will not tell you about the work that matters most. The final pillar of the AI Adoption Series argues that adoption stalls on trust, not technology: when people hide how they use AI, the expertise stays hidden with them. Includes what transparency and inclusion look like in practice, from written commitments to who gets a say in the rollout.

At SAP Sapphire in Orlando, CEO Christian Klein unveiled the Autonomous Enterprise, SAP's bold new vision for how global businesses will operate. The SAP Business AI Platform is a unified system for building, contextualizing, and governing AI agents across core business processes including supply chain, finance, HR, and procurement. Klein highlighted a structural gap in current AI adoption: most AI sits above the waterline trained on public data, while real value requires AI that understands mission-critical business data, end-to-end processes, and governance requirements. SAP is deepening partnerships with Anthropic, AWS, Google Cloud, Microsoft, NVIDIA, and Palantir to close that gap. For business leaders, the signal is clear: the next phase of enterprise AI is not about pilots. It is about agents running the business.

Microsoft repositioned Foundry as the operating environment for AI agents in production rather than another model catalog, with more than 11,000 models under one control plane for moving agents from prototypes to production.

As AI agents move from experimental tools to active participants in business operations, traditional governance frameworks are proving inadequate. Unlike standard software, AI agents can make autonomous decisions, access sensitive systems, and interact with other agents, creating risks that static rules and periodic audits cannot catch in time. Organizations are being urged to build governance structures designed from the ground up for agentic AI. For business leaders, the central message is that deploying AI agents without agent-specific oversight is a material risk, not just a technical gap. Companies that get governance right early will be better positioned to scale AI safely and maintain stakeholder trust.

Chinese tech workers are increasingly creating AI models trained on their own work styles, communication patterns, and decision-making habits, essentially digital stand-ins capable of handling routine tasks on their behalf. This trend is emerging partly as a survival strategy, as workers face intense pressure to prove their value in workplaces rapidly adopting automation. At the same time, many of these workers are resisting how their employers deploy AI, raising concerns about surveillance, data ownership, and the boundaries of human-AI collaboration. For business leaders, this signals that AI adoption is not simply a technology rollout. It is a deeply human negotiation that touches employee identity, trust, and autonomy. Organizations that fail to engage workers as stakeholders in AI deployment risk quiet resistance, disengagement, or talent attrition. Building clear policies around data use, worker consent, and the role of AI in performance evaluation will be essential to sustaining both adoption and morale.

Morgan Stanley released a sweeping report warning that a transformative leap in artificial intelligence is imminent, driven by unprecedented levels of compute at America's top AI labs. The investment bank specifically cited OpenAI's GPT-5.4 "Thinking" model, which scored 83% on the GDPVal benchmark, placing it at or above the level of human experts on economically valuable tasks. Executives at major AI labs are telling investors to expect progress that will "shock" them, and Gartner projects worldwide AI spending will hit $2.52 trillion in 2026 alone. The report's central message: most enterprises are still not structurally prepared to capture the value that the next wave of AI will make available.