The Pyramid Breaks
Tim Gonzalez@timgonz10
+5.2%
since Oct 6
The idea
I believe the law firm pyramid is breaking: AI handles the associate work, so leverage stops meaning headcount and value shifts to judgment, data, and outcome-based fees. My spearhead owns that directly through legal data and AI platforms, litigation finance, and e-discovery. Behind it sits the same repricing across white-collar work: software charging for work done, the institutions keeping the margin, the data moats underneath, and a small sleeve of AI-native challengers.
What it owns
- As AI takes over junior legal work, the companies selling the tools and outcome-priced economics that replace it capture the value. · Thomson ReutersIt owns Westlaw, half the legal research duopoly, and when AI tools replace associates doing research, those tools still license its case law and citations., RELX plcIt owns Lexis, the other half of the legal data duopoly, and as AI becomes how lawyers use research, that data still routes revenue to it., Burford CapitalIt funds litigation and gets paid on results rather than hours, and AI cutting the cost of running cases directly widens its returns., IntappAs law firms restructure with fewer associates and new fee arrangements, they buy its software to manage pricing, deals, and compliance instead of headcount., CS DiscoIt sells AI document review directly to law firms, winning if they replace review teams with its software rather than building their own.28%
- Platforms are shifting from charging per user to charging for work their AI completes, repricing white-collar labor across industries. · MicrosoftIt is the default vendor where enterprises buy AI agents through Copilot and Azure, so it captures fees whenever professional work shifts from people to software., ServiceNowIt already runs workflows inside enterprises and now offers agents that complete the work those workflows used to route to people, pricing by work done instead of seats., SalesforceIts Agentforce charges per conversation resolved, applying outcome pricing to sales and service work and showing that agencies adoption shows up in results., IntuitIts agents perform the work accountants once billed for, while the company keeps the fee, as long as small businesses buy the agent instead of hiring humans.23%
- Banks and insurers spend enormously on white-collar labor, and the margin gains from automating it show up in their bottom line. · JPMorgan ChaseIt employs more white-collar workers than most competitors and is deploying AI against that cost base in legal, compliance, and core operations to improve margin., Goldman SachsIts cost base is almost entirely expensive professionals doing the work AI reprices, and junior analyst work moving to software drops straight to its bottom line., ProgressiveIt already runs leaner on technology than peers, and claims and underwriting are dense with the professional work AI reprices, letting it convert automation to margin faster., American ExpressIt runs huge customer service and back-office operations that AI agents now handle, and as a buyer rather than seller of automation it keeps the savings as margin.20%
- AI agents are only as good as the data they run on, and these firms own financial and insurance data that cannot be replicated or scraped. · S&P GlobalIt owns ratings, indices, and market data that any AI doing financial analysis must license, so agents multiply demand for its feeds., Verisk AnalyticsIt owns decades of claims and risk data that AI underwriting and claims agents require, making its data the standard insurers automate white-collar work around.11%
- Companies built with AI doing the labor from inception never carried the payroll burden that incumbent competitors must now shrink. · LemonadeIt runs insurance with AI handling what underwriters and adjusters did, proving that starting without payroll beats cutting it, and its loss ratios show whether the model holds., DuolingoIt makes with AI what used to need teams of content creators and teachers, showing that a company built without white-collar payroll can grow users through AI-generated content., Upstart HoldingsIt replaces human loan underwriters with AI models, pricing white-collar judgment work by results, though it swings with credit conditions and only works through a full cycle.8%
- Cash10%