AI's squeeze on finance: stocks most at risk, from Schwab to S&P Global

AI's Next Test For Financials: Who Feels The Squeeze?

AI is moving into advice, underwriting, research, and customer service. The near-term benefit is cost relief. The longer-term risk is pricing pressure in products and services that start to look like commodities.

For investors and operators, the question isn't "if" AI changes the model - it's where profits compress first and who has the moats to defend them.

Most Exposed Segments (And Why)

• Wealth managers and platforms (SCHW, RJF, MS): AI advisors and model portfolios push fees down. Client service, planning, and basic portfolio construction get automated, reducing the perceived gap between premium advice and low-cost digital options.
• Insurance brokerages (Aon, Willis Towers Watson, Arthur J. Gallagher, Marsh McLennan): AI-assisted risk discovery, quoting, and benchmarking compress placement time and reveal pricing more transparently. The edge shifts to proprietary data, specialized lines, and complex risk structuring.
• Data-centric providers (SPGI, NDAQ): Public information is easier to aggregate and analyze with LLMs, pressuring data packages built on widely available sources. Proprietary datasets, ratings IP, and index licensing stay defensible; generic feeds face pushback on price.
• Smaller banks: Digital origination, AI chat, credit models, and fraud tools require scale. Without it, customer acquisition costs stay high while deposit pricing and fee income face pressure from AI-smart competitors.

Margin Mechanics To Watch

• Wealth management take rates: Expect gradual fee compression as AI planning tools and hybrid advice become default. Monitor advisory fee bps, revenue per advisor, and client migration to lower-cost tiers.
• Cash economics: As clients use AI to optimize cash, sweep balances and NII tailwinds normalize. Track sweep balances, deposit betas, and money market mix.
• Brokerage placement yields: Faster quoting and more transparent comps narrow spreads. Look for changes in placement commissions and retention in complex lines.
• Data subscription pricing: Buyers will unbundle and renegotiate where content is replicable with public sources. Watch churn, ARPU, and attach rates for proprietary modules.

Where The Moats Still Hold

• Proprietary data and IP: Ratings methodologies, hard-to-replicate alt data, index licensing, and long time-series remain sticky.
• Complexity and relationships: Large commercial risk, bespoke solutions, private markets, and high-touch wealth keep human-led value.
• Distribution and ecosystem: Custody networks, advisor platforms, liquidity pools, and multi-product bundles make switching costly.
• Scale for model training: Firms training on proprietary client interactions and outcomes can compound service quality and cost advantages.

Company-Level Red Flags

• High revenue mix from standardized, public-data products.
• Advisory fees clustered near premium pricing with weak differentiation.
• Heavy reliance on cash sweep economics or payment for order flow.
• Elevated cost-to-income with limited AI automation roadmap.
• Low product depth and thin distribution vs. larger peers.

Practical Moves For Operators

• Automate the "middle 60%" of work: Onboarding, KYC, servicing, research drafts, model updates, and claims triage. See the AI Learning Path for Administrative Assistants for role-specific upskilling and workflow examples.
• Reprice before you're forced: Introduce tiered advice and modular data bundles to defend ARPU while expanding reach.
• Lock in proprietary advantage: Build or acquire unique datasets, audit trails, and outcomes data that feed your models.
• Re-skill the front line: Advisors and brokers who use AI copilots will outproduce peers on both service and wallet share.

Practical Moves For Investors

• Stress-test fee take rates down 10-30 bps for wealth managers; test 5-10% price giveback on commoditized data lines.
• Model sweep balance normalization and higher deposit betas through a full rate cycle.
• Segment revenue into proprietary vs. public-data exposed; assign different durability haircuts.
• Track AI opex savings and redeployment: cost-out is table stakes; reinvestment into data and distribution sets the winners.

Names Frequently Cited As At-Risk (Exposure Varies)

• Wealth platforms: Charles Schwab (SCHW), Raymond James (RJF), Morgan Stanley (MS)
• Insurance brokers: Aon (AON), Willis Towers Watson (WTW), Arthur J. Gallagher (AJG), Marsh McLennan
• Data/Exchanges: S&P Global (SPGI), Nasdaq (NDAQ)

Risk is not uniform. Firms with deeper proprietary data, complex product mix, and strong distribution can offset fee pressure with scale, cross-sell, and AI-driven productivity.

Further Reading

• BIS: AI and machine learning in financial services
• McKinsey: Generative AI's impact on banking

Upskilling Your Team

If you're building internal competency around AI for research, risk, and client service, a curated view of tools helps. See a practical roundup here: AI tools for finance. For structured training focused on research workflows, consider the AI Learning Path for Research Associates.