Firms Deploying AI at Scale Are Capturing Efficiency as Margin — Not Passing It to Clients

The Legal Stack's Legal AI Fee Arrangement Disruption Report 2026 — based on interviews with 14 Fortune 500 legal ops directors, 23 AmLaw 200 billing partners, and analysts at Burford, Omni Bridgeway, and Deminor — documents a clear pattern: firms with AI deployed at scale are not passing the majority of efficiency gains to clients through lower fixed fees, but rather retaining them as margin, faster delivery, or competitive differentiation. In M&A diligence, AI-assisted contract review at Simpson Thacher, Kirkland & Ellis, and Latham & Watkins cut junior associate hours by 30–40% on qualifying matters — with neither firm reducing its blended rate nor moving to fixed-fee diligence at meaningful scale. Real estate transactional work is the exception: fixed-fee commercial lease review now accounts for over 60% of that revenue at Goulston & Storrs and Seyfarth Shaw, up from roughly 40% pre-2024, with margin increasing as AI-assisted abstraction reduced paralegal hours while clients accepted project pricing they already preferred. The emerging hybrid structure — a base fee slightly below historical hourly equivalents, a "complexity collar," and quarterly AI compute-cost disclosure to clients — has no industry name, is negotiated matter-by-matter, and is available primarily to sophisticated Fortune 500 legal operations teams.