AI agents could strip $170B from bank profits - here's the finance playbook
Analysts warn that banks may forfeit up to $170 billion in profits as autonomous financial agents push customers toward higher-yield options. McKinsey analysts note these agents can scan markets, compare rates, and nudge customers to move idle cash out of low-interest accounts, compressing margins across the system.
Research cited in the market shows customers hold about $23 trillion of $70 trillion in assets in minimal-interest accounts. If AI agents start managing these balances at scale, sector profits could fall by roughly 9%. Banks expect their own AI to save up to 20% in costs, but competitive pressure is likely to pass much of that value to consumers.
Why this matters for P&L
- Funding costs rise: Deposit beta increases as customers reprice faster. Low-cost, non-interest-bearing balances shrink.
- NIM pressure: More switching to money market funds, T-bill ladders, and high-yield savings compresses spread income.
- Retention spend goes up: Promotions, targeted rate bumps, and rewards dilute margin to keep balances in-house.
- Fee mix shifts: Flows move toward brokers and custodians that package automated cash optimization.
How AI agents change deposit behavior
Friction to switch drops to almost zero. An agent can open, fund, and sweep across institutions in minutes, monitoring safety, yield, and liquidity in real time. That makes "sticky" low-rate deposits less sticky, especially for rate-sensitive segments like SMBs, affluent retail, and corporate treasuries.
12-24 month scenario view
- Base case: Gradual adoption; 3-5% NIM impact in the most exposed portfolios, profit headwind near the cited 9% if agent usage scales.
- Bear case: Fast agent uptake among HNW and SMB segments; accelerated outflows to brokered cash and MMFs.
- Bull case: Banks deploy their own agents and sweep platforms, recapturing flows into CDs, T-bills, and in-house funds.
What finance leaders should do now
- Reprice proactively: Model deposit beta by segment; stagger targeted rate increases before outflows force broad hikes.
- Stand up first-party agents: Offer customers automated cash management with rules for liquidity, risk, and tax.
- Harden the deposit mix: Promote CDs, insured cash sweep, and treasury products to anchor balances.
- Automate retention: Real-time triggers for rate-matching and personalized offers when agent-like behavior is detected.
- Broaden product shelf: Easy paths to T-bills, short-duration funds, and sweep programs inside your ecosystem.
- Strengthen liquidity buffers: Update stress assumptions for higher churn and faster repricing cycles.
- Revise transfer pricing: Reflect true optionality of deposits in FTP and performance measurement.
- Compliance and data guardrails: Clear consent, API security, and audit trails for agent-driven actions.
Metrics to watch
- Deposit beta by segment and product
- Rate-sensitive balance share and churn to MMFs/brokerage
- Time-to-reprice and offer acceptance rates
- Agent-originated account openings and sweeps (internal vs external)
- Liquidity coverage and cost-to-income ratio
Signals from the market
- Consumer apps adding "auto-move my cash" features across banks and brokers
- Broker custodians marketing agent-based cash optimization to SMBs
- Banks piloting AI-driven treasury and cash sweep tools for retail and business clients
Banks will try to offset the hit with their own AI and ops savings. Expect competition to pass much of that gain to customers until pricing stabilizes. The winners will be the institutions that make switching less attractive by offering smart automation, transparent yields, and immediate liquidity inside their walls.
For context on how AI is reshaping banking economics, see analysis from McKinsey. If you're building internal capability, this catalog of finance-focused tools is a practical starting point: AI tools for finance.
Your membership also unlocks:
