Insurers are embedding AI into underwriting, the function that decides which risks to take and at what price, per the 2026 Evident AI Index, an annual benchmark of 30 large insurers run by research firm Evident.
Across the 30 insurers tracked by the 2026 Evident AI Index, an annual benchmark run by research firm Evident, AI-specialist headcount grew 32% over the past year while the broader insurance workforce contracted 2.2%. That contrast is the real story: not "AI is coming for insurance jobs" but "AI is reshaping who insurers hire," and where those new hires are directing their work.
Underwriting is the epicenter. Underwriting is the function where insurers decide which risks to take on and at what price, and the 2026 Evident AI Index reports that insurers are now embedding AI into workflows that directly influence underwriting discipline and capital allocation, not just customer service or claims triage. The shift puts AI in the room where pricing and risk selection are set, the core capital decisions of the business.
A second structural signal sits behind the hiring numbers. Nearly 40% of the indexed insurers now have a senior leader with explicit AI responsibility, and most of those appointments happened in the last 12 months. That is a new C-suite tier that did not exist at this scale a year ago. Christian Preece, Insurance Director at Evident, frames the shift as a move from "competing on AI ambition" to "competing on AI value creation," according to a recent industry analysis of the Index.
The ROI-disclosure push is partly a response to pressure. Preece is explicit that rising shareholder and board scrutiny of AI cost versus value is forcing insurers to translate AI work into financial outcomes tied to risk, not to story-time metrics. Read that as accountability architecture, not maturity self-congratulation: a regulated, risk-heavy industry under investor pressure to convert AI spend into capital discipline. A vendor benchmark that calls for hard ROI disclosure is, in this case, riding the same wave that shareholders and directors are already pushing.
A forward signal sits underneath the hiring data. Roughly one in four newly disclosed AI use cases in the Index now involves agentic AI, meaning AI systems that coordinate multi-step tasks (such as claims handling or policy administration) with less human handoff. That is still a minority share, but it is the fastest-growing slice of the disclosed pipeline, and underwriting is the most common front door. The point is not that agentic AI has arrived at scale; it is that leading insurers are willing to disclose the use case at all, which is a different posture than 12 months ago.
The 2.2% workforce contraction is a real human cost, not a footnote. It reflects a workforce being reshaped: more AI specialists, fewer staff in roles that AI can now perform. The transferability for other regulated industries is the pattern, not the headcount: name a senior AI owner, measure hard ROI, accept that some roles will shrink while AI-skilled roles expand, and disclose both.
What to watch next: whether the same insurers extend ROI disclosure specifically to underwriting decisions, where pricing accuracy and risk selection are easier to defend to regulators than to shareholders. If they do, expect the next round of disclosure to look less like a marketing claim and more like a capital filing.