AI Insights Feb 20 2026

Feb 20, 2026

insuranceindustry.ai article hero image

AI Insights: February 20, 2026

Your weekly analysis of AI developments in insurance.

Travelers Turns Its AI Strategy Into a Live Voice Agent

Last week we reported on Travelers deploying AI across 20,000 employees and hinting at leaner claims call centers. Now we know what that looks like in practice. On February 18, Travelers launched AI Claim Assistant, a fully agentic voice service built on OpenAI’s Realtime API that handles auto damage claim calls from start to finish, without a human on the line.

The system doesn’t just collect information. It consults with customers, provides policy details, helps them decide whether to file, and then transitions them to a digital experience where they can upload photos, schedule repairs, and arrange a rental car. Travelers Chief Claim Officer Nick Seminara said early customer feedback has been “overwhelmingly positive.” OpenAI’s Head of Platform Product called it “one of the most sophisticated agentic voice implementations” built on their platform.

Call center employees are being retrained and repositioned, not simply let go. Travelers is framing this as an upskilling program that continues as the technology expands to additional lines of business.

Why This Matters:

Travelers has now publicly closed the loop on a strategy that many carriers are still in pilot mode on. In January, it equipped 10,000 engineers with Anthropic AI assistants. In Q4 earnings, it reported 20,000 employees using AI tools regularly. Now it’s put an autonomous voice agent on the phone with real customers filing real claims. Each step builds on the last.

This is what enterprise-scale AI deployment looks like when a carrier commits to it. The $1.5 billion Travelers invested in AI and technology in 2025 alone isn’t abstract spending. It’s showing up as a voice system that handles the full complexity of a claim conversation at scale.

Strategic Implications:

For carriers still evaluating agentic AI for claims, Travelers just made the competitive calculus more urgent. Customers who file through Travelers’ AI Claim Assistant get immediate guidance, consistent service, and a seamless handoff to digital tools. That’s a different customer experience than waiting on hold.

The agentic pattern here is worth noting: this isn’t a chatbot that deflects simple questions. It’s an orchestrated system that consults, advises, collects information, and triggers downstream actions. Every carrier with a claims call center should be asking how long before their customers notice the difference.

ERGO Puts a Number on What AI Costs in Jobs

Munich Re’s primary insurance unit ERGO announced February 17 that it will eliminate approximately 1,000 positions in Germany by 2030, citing increased use of artificial intelligence. The cuts target telephony and claims processing, the same functions Travelers just automated with its voice agent. ERGO plans to reduce headcount by about 200 positions per year through natural attrition, early retirement, and voluntary programs. Forced redundancies are off the table, and 500 employees will be retrained for roles in growth areas like retirement planning.

ERGO employs around 15,000 people in Germany. The cuts align with Munich Re’s broader goal of reaching approximately €600 million in annual cost savings by 2030.

Why This Matters:

The pattern emerging across the industry is consistent enough now to be a trend, not a series of isolated decisions. Accenture laid off 11,000 earlier this year tied to AI restructuring. Acrisure cut 400 accounting jobs. Allianz Partners cut 1,800 positions in late 2025. Now ERGO is adding 1,000 more to that count.

The jobs being eliminated share common characteristics: routine, repetitive, process-driven tasks in telephony and back-office functions. These are precisely the workflows agentic AI handles best. The ERGO announcement is notable not because it’s surprising, but because it’s becoming routine. That’s a signal worth paying attention to.

The 500-person retraining commitment is meaningful. It suggests ERGO recognizes that managed transitions produce better outcomes, for the organization and for the people inside it, than abrupt cuts. How well that retraining program works will tell us a great deal about what “responsible AI adoption” actually delivers in practice.

Strategic Implications:

For insurance executives managing workforce decisions, ERGO’s approach provides a useful model: a five-year runway, no forced redundancies, parallel investment in retraining, and focus on functions where AI genuinely has the advantage. That’s a defensible framework for a difficult conversation.

For carriers and agencies not yet facing this question, the ERGO announcement is a preview. The functions AI is replacing at ERGO (inbound calls, claims intake, routine processing) exist in virtually every insurance operation. The question isn’t whether this happens in your organization. It’s when and whether you’ve built the plan to manage it thoughtfully.

Lockton Re: AI Needs Its Own Risk Category

A new report from Lockton Re, produced with Lockton International and Armilla AI, is making the case that AI has outgrown existing insurance classifications. The report, released February 18, argues that current commercial lines policies were not designed for AI risk and that the industry is developing a growing gap between what policies intend to cover and what they actually cover when AI is involved.

The report maps AI exposures across commercial classes and finds coverage that is silent, fragmented, or misaligned with how AI failures actually happen. Algorithmic decision errors, hallucinations, misguidance, and data training issues are showing up as named causes of loss in some endorsements, but Lockton Re warns these definitions are often too narrow. When an incident falls outside the specific named peril, coverage may not respond.

The systemic risk finding deserves particular attention. When a widely used foundation model contains flawed training data or an unintended performance characteristic, failures can occur simultaneously across organizations in different industries and geographies. Armilla AI co-founder Baiju Devani put it directly: the question for the insurance industry isn’t whether AI will create systemic risk events, but when, and whether underwriting practices can keep pace.

Why This Matters:

The AI exclusion trend we covered last week is one side of this story. Carriers pulling back from AI-related liability is a defensive response to a risk they can’t adequately price. The Lockton Re report frames the other side: the risk isn’t going away, and someone needs to underwrite it properly.

Commercial general liability carriers are currently not modeling, underwriting, or pricing AI risk. The CGL form was written for a different world. As AI is embedded deeper into operations, product decisions, and customer interactions across insured businesses, the exposure grows while the coverage framework stays static. That’s how coverage disputes get expensive.

Strategic Implications:

For risk managers and commercial clients, the immediate action is gap analysis. Where in your operation does AI touch decisions, customer interactions, or products? Map the exposure, then examine your policy language for how an AI-related failure would actually be treated at claim time. Don’t assume your technology E&O covers it. Many endorsements cover developers of AI solutions, not the companies using them.

For carriers and MGAs, the Lockton Re report is a roadmap for a market that needs to develop. Fewer than five meaningful AI-specific insurance products currently exist, according to industry experts. A carrier that builds genuine underwriting capability for AI risk, rather than writing blanket exclusions, is positioning for a market with no current ceiling.

Health Insurers Bet $1 Billion on AI While Providers Sound Alarms

During Q4 2025 earnings calls, health insurer executives across the industry were nearly uniform in their AI messaging. UnitedHealth Group CEO Stephen Hemsley declared the company is “clearly embarking on a new age of technology” and pledged to invest $1.5 billion in AI in 2026. UnitedHealthcare CEO Timothy Noel projected AI-enabled operating cost reductions of nearly $1 billion this year. Over 80% of member calls already use AI tools to route and respond to questions.

At the same time, a February 17 report from STAT News documented growing friction between health insurers and providers, with healthcare systems expressing concern that AI-driven efficiency is coming at the cost of care access. The backdrop includes UnitedHealth’s ongoing litigation over its nH Predict algorithm, which was alleged to have denied Medicare Advantage claims with a 90% error rate, and a Department of Justice investigation into Medicare billing practices.

UnitedHealth has since overhauled its responsible AI framework, established a review board with clinicians and ethicists, and is publishing prior authorization rates and approval statistics as part of a transparency push. The question the industry is watching is whether that framework changes outcomes, or just the optics.

Why This Matters:

This story matters to P&C executives as much as health insurance leaders, because the dynamic playing out in health insurance is the leading edge of what happens when AI cost-reduction goals collide with regulatory oversight and public trust.

The numbers driving UnitedHealth’s AI investment are straightforward: AI handles volume at a cost that human call center staff cannot match. But when AI-driven efficiency translates into claim denials that patients can’t effectively appeal, the regulatory and litigation consequences dwarf the savings. UnitedHealth’s experience with nH Predict is a case study in what happens when the efficiency case for AI isn’t matched by adequate governance.

Strategic Implications:

For P&C carriers deploying AI in claims, underwriting, or customer service, the UnitedHealth situation is a governance blueprint, read in reverse. The risks that materialized at UnitedHealth (error-prone automated decisions, inadequate appeals processes, regulatory scrutiny, class-action litigation) all have direct analogues in commercial and personal lines operations.

The carriers that will get this right are those treating AI governance as a competitive asset, not a compliance cost. Travelers’ approach of pairing AI deployment with visible workforce transition planning and public statements about human oversight is the model. Deploying AI for efficiency while building in explainability, audit trails, and genuine human review at decision points protects both customers and the organization.

Insurtech Investment Hits $5 Billion, But ROI Remains Elusive

Gallagher Re’s Q4 2025 Global InsurTech Report, published February 16, puts the first concrete number on what insurtech attracted last year: $5.08 billion, up 19.5% from 2024. Two-thirds of that capital went to AI-focused companies. The Q4 surge alone reached $1.68 billion across roughly 230 AI deals, the highest quarterly total since 2022.

Property and casualty insurtechs led the rebound, with funding up 34.9% to $3.49 billion.

The report is careful to note what the money hasn’t yet resolved: the ROI paradox. Companies are achieving efficiency gains, freeing up employee time and automating tasks, but many haven’t found where the newly available resources create proportional revenue gains. Efficiency without productivity improvement produces cost reduction, not transformation.

Gallagher Re also flagged that AI company valuations have significantly outpaced demonstrated revenue generation, raising questions similar to those that preceded the dot-com correction.

Why This Matters:

The $5 billion figure tells you the industry believes AI is real. The ROI paradox tells you execution is harder than investment. Both things can be true simultaneously, and understanding that tension is essential for executives making decisions about where to put resources.

The distinction Gallagher Re draws, evaluating AI at the product level, the company level, and the industry level separately, is analytically useful. An individual carrier can capture real efficiency gains from AI even if industry-level transformation is slower than the hype suggests. The question is whether those efficiency gains compound into competitive advantage or simply reset the cost baseline for everyone.

Strategic Implications:

For carriers and agencies evaluating AI investments, Gallagher Re’s framework is practical. Start with the product level: does this specific AI application solve a specific operational problem with measurable results? Don’t buy the transformation narrative before you’ve validated the transaction.

The insurtech funding rebound means the vendor landscape is growing, not contracting. More options arriving with more capital behind them creates better buying conditions for insurers, but also more noise to filter. Organizations with clear criteria for what AI needs to actually deliver, not just promise, will make better decisions in this environment than those buying the category.

The Bottom Line

This week’s stories connect more directly than they might appear. Travelers deployed a live agentic voice system. ERGO announced 1,000 jobs eliminated by the same technology. Lockton Re documented the coverage gaps that AI creates in commercial insurance. UnitedHealth revealed the governance risks when AI cost-cutting runs ahead of oversight. And Gallagher Re confirmed the money is real, even if the ROI proof points are still developing.

The thread running through all of it is execution. Carriers that are winning with AI right now are the ones that have moved from strategy to deployment, built governance alongside capability, and managed the workforce dimension with intention rather than avoiding it.

The insurtech investment surge means the tools are available. The ERGO and UnitedHealth stories mean the stakes of getting implementation wrong are clear. The Travelers story means the benchmark for what good looks like is already in the field.

The gap between what AI can do and what most organizations are actually doing with it is narrowing fast. The carriers closing that gap are the ones that will set the competitive standard everyone else responds to.

AI Insights appears every Friday, analyzing AI developments through an insurance lens. For deeper analysis of strategic implications, visit InsuranceIndustry.ai.

By James W. Moore

Sources:

Subscribe to AI Insights Newsletter

AI Disclaimer: This blog post was created with assistance from artificial intelligence technology. While the content is based on factual information from the source material, readers should verify all details, pricing, and features directly with the respective AI tool providers before making business decisions. AI-generated content may not reflect the most current information, and individual results may vary. Always conduct your own research and due diligence before relying on information contained on this site. 

Related Posts

468