AIG Completes Its Metamorphosis: A Pure-Play P&C Giant Betting Hard on Agentic AI

AIG has shed Corebridge, installed a new CEO from Aon, and is deploying multi-agent AI systems across underwriting and claims. With a 88.1% accident year combined ratio and $7.5B in quarterly net premiums written, the transformed insurer is now a focused commercial P&C machine trading at just 1.0x book.

AIG · Financials · September 05, 2026

S&P 500 Position

AIG sits in the mid-tier of S&P 500 Financials. Within multi-line and P&C insurance, it is smaller than Chubb ($110B+), Progressive ($150B+), and Travelers (~$60B) by market cap but remains one of the largest commercial P&C underwriters globally by premium volume. Its $39.8B market cap is roughly half of Chubb's, reflecting years of capital return rather than growth-driven valuation expansion.

Index Weight: ~0.08% | Rank: Approximately #250-280 in the S&P 500 by market cap

Company Overview

AIG in September 2026 is a fundamentally different company than the sprawling conglomerate that nearly collapsed the global financial system in 2008. The final sale of its remaining Corebridge Financial stake in May 2026 completed a five-year separation process, leaving AIG as a pure-play global commercial property & casualty insurer operating across 80+ countries. The company writes $7.5 billion in net premiums per quarter across North America Commercial, International Commercial, and Global Personal lines, competing head-to-head with Chubb, Travelers, Zurich, and AXA XL for large, complex multinational risks. The leadership transition is equally stark. Peter Zaffino, who delivered five consecutive years of underwriting profitability and an 85%+ share price increase, steps down as Executive Chair on September 15 and will join Palantir in January 2027. Eric Andersen, who grew Aon's market value from $35 billion to $85 billion as its President, took over as CEO on June 1. Andersen inherits a company with a clean balance sheet (0.22x debt/equity), a $10 billion buyback authorization, and a deeply embedded AI infrastructure strategy that is already producing measurable underwriting gains. What makes AIG technically interesting right now is the 'AIG Next' digital transformation program. The company has sunset 1,200 legacy applications, built an enterprise ontology (a digital twin of its business processes and data), and is deploying agentic AI across underwriting workflows. This is not PowerPoint AI — Lexington's middle-market property business cut quote generation time by 55% and improved submit-to-bind ratios by 35% using these tools. AIG is constructing an orchestration layer to coordinate multiple AI agents across the enterprise, a genuinely ambitious architectural bet for a century-old insurer.

Products & Revenue

AIG's revenue is now almost entirely General Insurance premium income. After the Corebridge exit, investment income and residual life/retirement fees have shrunk dramatically. Q2 2026 General Insurance net premiums written totaled $7.52 billion, with North America Commercial as the largest contributor at $3.12 billion. Net investment income of $1.13 billion supplements premium revenue but declined sharply year-over-year as Corebridge-related investment assets rolled off the balance sheet.

North America Commercial (~41%): Large commercial and specialty P&C lines in the US and Canada — property, casualty, financial lines, Lexington surplus lines. Generated $3.12B in Q2 2026 revenue, up 9% YoY. Net premiums earned of $2.32B.

International Commercial (~30%): Commercial P&C across Europe, Asia-Pacific, and emerging markets. Includes multinational programs, specialty lines, and Lloyd's syndicate participation. Net premiums earned of $2.27B in Q2 2026, up 7% YoY.

Global Personal (~14%): Personal insurance including travel, accident & health, and auto/homeowners in select international markets. Profitable but smaller than the commercial books.

Net Investment Income (~15%): Income from the General Insurance investment portfolio. Totaled $1.13B in Q2 2026, down significantly from $1.47B a year earlier as Corebridge-related assets exited.

Based on AIG Q2 2026 10-Q filing and earnings release (August 6, 2026). Revenue percentages approximated from segment-level data; General Insurance now constitutes substantially all of AIG's operations post-Corebridge exit.

Leadership

Eric Andersen

CEO since 2026. Became President and CEO on June 1, 2026. Previously served as President of Aon from 2020 to 2025, where he oversaw the brokerage's growth from $35 billion to $85 billion in market value. His deep broker-side relationships and understanding of risk placement economics give him a distinctive vantage point for running a carrier.

John Rice, Incoming Board Chair (effective September 15, 2026): Succeeds Peter Zaffino as Board Chair. Previously a Vice Chairman of GE and CEO of GE Global Growth and Operations. Brings large-scale operational transformation experience.

Scott Hallworth, Chief Digital Officer: Joined September 2025 from HP, where he was Chief Data & Analytics Officer. Oversees AIG's digital, data, and generative AI strategy. Responsible for scaling 'Underwriting by AIG Assist' and the enterprise ontology/digital twin initiative.

Nancy Bewlay, EVP & Global Chief Underwriting Officer (effective September 8, 2026): Reports directly to CEO Andersen. Leads the global underwriting organization at a critical moment as AI-assisted tooling reshapes submission processing and risk selection.

Christine Williams, Head of Global Client and Broker Relationships (effective September 1, 2026): New role created to centralize AIG's distribution strategy across brokers and direct channels, signaling Andersen's intent to deepen client-facing capabilities.

The AI Angle

Agentic AI Underwriting at Industrial Scale

AIG's AI strategy is the most technically ambitious in the P&C insurance industry. The company has shipped two core products: 'Underwriting by AIG Assist' and 'Claims by AIG Assist,' both powered by partnerships with Palantir (ontology and orchestration), Anthropic (LLMs), AWS (cloud infrastructure), and Google (additional model access). These are not chatbot wrappers — AIG is building an orchestration layer that coordinates multiple AI agents across the enterprise, each handling discrete tasks in the underwriting and claims workflows. The measurable results are striking. At Lexington, AIG's surplus lines subsidiary, AI-driven automation cut quote generation time by 55%, increased quoting submissions by 30%, and improved the submit-to-bind conversion ratio by 35%. Across the broader enterprise, AIG has processed over 370,000 submissions through the AI-assisted pipeline, tracking toward a 500,000-submission target by 2030. The tool has been rolled out to seven additional lines beyond its initial deployment. Architecturally, AIG is constructing what it calls an 'AIG ontology' — a digital twin of its entire business that maps concepts, processes, data elements, and workflows into a shared framework. This is directly derived from the Palantir Foundry/AIP approach and serves as the connective tissue between AI agents, legacy systems, and human underwriters. CDO Scott Hallworth, recruited from HP in September 2025, is driving this buildout after AIG sunset 1,200 legacy applications under the 'AIG Next' program. The competitive risk is execution speed. Every major commercial insurer is investing in AI-assisted underwriting — Chubb, Travelers, and Zurich all have active programs. AIG's advantage is its deeper infrastructure partnership with Palantir (whose incoming executive, Zaffino, architected the AIG strategy) and the structural willingness to eliminate legacy tech debt at scale. The departure of Zaffino to Palantir in January 2027 creates both continuity risk and a tighter strategic link between AIG and its primary AI platform partner.

Financial Snapshot

Revenue (TTM): $26.7B — TTM ending June 30, 2026 | Net Income: $3.0B net income TTM

Margins: Net margin 11.1%; combined ratio 88.1% (accident year, Q2 2026); General Insurance H1 2026 underwriting income $1.46B vs. $869M in H1 2025

AIG is cheap on every standard metric. Trading at book value with a 13.9x P/E signals that the market has not yet re-rated AIG for its post-transformation earnings quality. The $10 billion buyback authorization and $6.8 billion returned to shareholders in 2025 alone demonstrate aggressive capital return. The strategic investments — $2.1B for 35% of Convex (specialty reinsurer) and $642M for 9.9% of Onex — indicate AIG is deploying excess capital into fee-generating, capital-light adjacencies rather than pure organic growth.

1-Year Performance

$76.21, down 1.5% YoY — essentially flat in a year where the broader market has moved higher

AIG's flat performance reflects a tug-of-war between improving fundamentals (10% EPS growth, sub-89% combined ratio) and the market's uncertainty around leadership transition and the loss of Corebridge investment income. The stock briefly traded near $78 after the Q2 earnings beat but has drifted lower. The consensus $89 price target implies ~17% upside, and Goldman Sachs' March upgrade to Buy specifically cited industry-leading earnings growth prospects.

Recent News

Fun Fact: AIG's 'AIG Next' digital transformation program eliminated 1,200 legacy applications from its technology stack — roughly half its entire application portfolio. The company is now building what it calls an 'AIG ontology,' a Palantir-powered digital twin that maps every concept, process, data element, and workflow across the enterprise into a unified graph. This ontology serves as the coordination layer for multiple AI agents that autonomously handle steps in underwriting and claims workflows — making AIG one of the first traditional insurers to deploy multi-agent orchestration in production, not just in proof-of-concept.