Some links on this page are affiliate links: if you buy through them we may earn a commission, at no extra cost to you.
McKinsey’s $170 billion warning is real, but it is not a prediction that banks will suddenly lose $170 billion in annual earnings. The consultancy’s scenario says global banking profit pools could be about 9% smaller—roughly $170 billion in 2030 dollars—if banks fail to adapt as customers use third-party AI agents to compare products, move deposits and manage credit.
For bank customers, the potential upside is better rates, lower fees and more convenient financial decisions. For banks, the threat is losing control of the customer relationship and being forced to compete away margins that currently depend partly on customer inertia.
What McKinsey actually estimates
In its Global Banking Annual Review 2025, McKinsey estimates that global banking profit pools could decline by approximately $170 billion, or 9%, if banks do not reposition themselves for agentic AI.
Recommended Free Tools
The calculation uses an estimated global banking profit pool of about $1.8 trillion in 2030. The $170 billion figure is expressed in 2030 dollars. McKinsey also cautions that using 2030 as the reference point does not mean the entire impact will necessarily occur by December 31, 2030. The report describes a longer-term scenario unfolding over roughly the next five to ten years, or the next decade or so.
#1 Best Overall
This is an industry-wide estimate, not a forecast for every bank. It refers to profit pools—the total profits available across banking products and institutions—not bank revenue, assets, stock-market value or the earnings of every individual bank.
What is agentic AI?
Generative AI usually creates content or recommendations in response to a prompt. An AI assistant might summarize a bank statement, explain a fee or suggest a savings product.
Agentic AI goes further. Within permissions set by the customer and the institution, an agent can reason through a task, retrieve information, use software tools and take actions with limited human intervention. In banking, that could mean identifying a higher-yield savings account, comparing credit cards, moving money, optimizing payment balances or executing another financial decision.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
McKinsey’s discussion of agentic AI in retail and small-business banking is not a claim that consumers are already widely delegating all of these decisions to autonomous software. The risk depends on adoption, permissions, regulation and whether agents can connect safely to financial accounts.
Why could AI reduce bank profits?
1. It could reduce customer inertia
Many customers keep money in an existing checking or savings account because switching is inconvenient. They may not regularly compare rates, move balances or renegotiate fees.
An independent AI agent could continuously compare available products and act on the customer’s instructions. McKinsey says that moving only 5% to 10% of checking balances to the highest-market rates could reduce total industry deposit profits by 20% or more, according to its analysis of the customer effect.
That does not mean every transfer would be immediate or frictionless. Account restrictions, fraud checks, settlement times, tax considerations and customer approval requirements could all slow or prevent a move. But the basic economic pressure is clear: banks may have to pay more to retain deposits or risk losing them.
The Tool Desk
Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →2. It could shift control of the customer relationship
Today, a bank often controls the digital experience through its website, app, marketing and product recommendations. If a customer instead asks an independent agent, “Where should I keep my emergency fund?” or “Which card is best for my spending?”, the agent may control discovery, comparison and eventually the transaction.
That could weaken brand loyalty, direct digital traffic, cross-selling and relationship pricing. A bank might still provide the account, loan or card, but the customer relationship could increasingly belong to the platform that makes the recommendation.
3. It could pressure deposit spreads
Banks generally benefit when they can fund themselves with relatively low-cost deposits and lend or invest those funds at higher returns. If agents make it easy for customers to seek the best available rate, banks may need to raise deposit yields or offer more incentives.
Higher rates can benefit savers, but they can also reduce the spread a bank earns between funding costs and asset yields. The result could be better returns for customers alongside lower margins for banks.
Free tools Windows power users keep installed
One-click scans. No signup required.
4. It could weaken credit-card economics
Credit-card customers may use agents to compare rewards, fees and balance-transfer offers, choose the cheapest borrowing option and optimize when or how much they repay. More active shopping could reduce interest income, fees and customer lock-in.
McKinsey’s January 2026 summary, “Banking’s AI angst,” identified credit-card lending as one of the most exposed businesses, with an estimated 34% decline in its profit pool in the relevant disruption scenario. Consumer deposits were estimated to face an approximately 27% decline. These are modeled scenario figures, not observed losses.
Which banking products are most exposed?
| Business | Why AI agents could matter | McKinsey scenario context |
|---|---|---|
| Consumer deposits | Agents could compare rates and move balances more easily. | Approximately 27% potential profit-pool decline. |
| Credit cards | Agents could compare rewards, fees and balance transfers and optimize repayments. | Approximately 34% potential profit-pool decline. |
| Payments and product distribution | Independent agents could control recommendations and transaction routing. | Exposure depends on adoption and platform control. |
| Mortgages | Agents could compare rates and improve shopping, although underwriting and transaction complexity create friction. | Expected to be less severely affected, but not insulated. |
| Wealth management | Agents could assist with product comparison and financial planning. | Expected to be less severely affected, but advice, trust and execution remain exposed. |
The figures in the table are McKinsey estimates for a particular disruption scenario. Product mix, geography, regulation and customer behavior will determine how individual institutions are affected.
The productivity paradox: AI can help banks and still hurt their margins
McKinsey does not argue that AI is uniformly negative for banks. Its analysis says agentic AI could reduce bank operating costs by 20% or more, equivalent to roughly 9% to 15% of operating profits.
Potential uses include customer service, credit underwriting, fraud detection, compliance, know-your-customer and anti-money-laundering work, software development, document review, operations and relationship-manager support.
The problem is who captures those savings. A bank may initially keep some of the benefit. But competitors could use lower costs to offer better deposit rates, cheaper credit or lower fees. Customers may receive more value, while technology platforms and AI-agent providers capture another share. In a competitive market, productivity improvements do not automatically become permanent bank profits.
This is why AI could improve efficiency inside a bank while reducing the long-term profitability of traditional banking products. The technology may make the system more productive, but competition can transfer that productivity to customers.
Which banks are best positioned?
A bank with an AI chatbot is not necessarily an AI leader. More meaningful signs include:
- AI deployed in production workflows rather than only in demonstrations.
- Accurate, accessible and permissioned data.
- Modern technology infrastructure that can connect AI to transaction systems.
- Strong identity, authorization, audit and monitoring controls.
- Measured improvements in costs, service, retention, conversion or risk outcomes.
- Governance that allows controlled deployment without bypassing financial safeguards.
- A strategy for remaining part of the customer’s decision and execution process.
McKinsey estimates that AI pioneers could gain as much as a four-percentage-point advantage in return on tangible equity over slower-moving institutions. Slow movers could face both high legacy costs and weaker customer relevance.
Smaller banks do not necessarily need to build their own frontier models. They may use governed third-party platforms or consortium services. That can reduce development costs, but it also creates integration, vendor-dependence and concentration risks.
Should banks build their own AI agents?
One response is for banks to create or embed their own financial agents. A useful bank-controlled agent could proactively identify a better deposit rate, explain a product trade-off, recommend a repayment strategy or execute a customer-authorized transfer.
The strategic tension is trust. An agent that always recommends the bank’s own products may be commercially convenient but less credible. An agent that genuinely optimizes for the customer could recommend a competitor and cannibalize the bank’s existing margins.
Quick wins for a faster PC:
Clear out junk files and repair common Windows errorsFree Scan →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Repair Windows errors before they cause bigger problemsFix Now →That trade-off may become unavoidable. Banks that refuse to help customers optimize may lose the relationship to independent agents. Banks that provide impartial optimization may preserve the relationship but give up some legacy pricing advantages.
Rank #4
Risks beyond profitability
AI agents that can act on financial accounts create risks distinct from the economic threat posed by third-party agents:
- Authorization: An agent could take an action beyond what the customer intended.
- Errors: Incorrect recommendations or outdated product information could cause financial harm.
- Fraud: Criminals could manipulate agents, credentials or connected tools.
- Prompt injection: Malicious instructions hidden in data or documents could alter an agent’s behavior.
- Privacy: Giving an agent access to financial information raises data-sharing and consent questions.
- Bias: Automated lending or product recommendations could produce unfair outcomes.
- Liability: Regulators and courts will need to determine responsibility when an agent makes a mistake.
- Explainability and records: Banks may need to document why an agent made a recommendation or executed a transaction.
- Vendor concentration: Dependence on a small number of model, cloud or platform providers could create operational risk.
These safeguards matter whether the agent belongs to a bank or an independent technology company. A recommendation-only system can be introduced with less risk than an agent permitted to move money, change repayment instructions or open an account.
What could make McKinsey’s estimate wrong?
The $170 billion scenario is conditional. The impact could be smaller if:
- Consumers adopt financial agents slowly or do not trust them with money.
- Regulation limits autonomous financial actions.
- Banks successfully retain the customer interface through their own agents and digital services.
- AI productivity gains remain with banks instead of being competed away.
- New AI-related revenue offsets pressure on deposits, cards and other traditional products.
- Technical, fraud or liability problems make automated switching less practical than expected.
The opposite is also possible: high-value customers could adopt agents before the mass market, making the economic impact significant even without universal use.
What this means for customers
Customers could benefit from easier rate shopping, more transparent fees, personalized recommendations and faster service. But convenience should not replace review. Before authorizing an agent to act, customers should understand:
- Which accounts and data it can access.
- Whether it can recommend actions, execute them or both.
- Transaction limits and approval requirements.
- How recommendations are paid for or influenced.
- How errors, unauthorized transfers and disputes are handled.
- Whether the agent compares the full market or only participating providers.
AI may make switching easier, but customers still need to consider account terms, withdrawal restrictions, taxes, deposit insurance, credit implications and the security of the service.
The bottom line
McKinsey’s warning is best understood as a warning about control of the banking relationship and future pricing power, not simply about job automation. AI could lower banks’ operating costs and improve service, yet third-party agents could simultaneously make deposits, cards and other products easier to compare and harder to profit from.
Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Clear out junk files and repair common Windows errors3Fix the driver behind crashes, sound loss and screen glitchesThe $170 billion figure is a conditional, modeled estimate of the potential reduction in global banking profit pools—not money banks have already lost or a guaranteed 2030 earnings decline. The banks most likely to fare well will be those that combine AI productivity with trusted, permissioned customer experiences while managing the risks of letting software make financial decisions.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
