BCG’s new Global Retail Banking Report identifies more than $370 billion in annual profit potential that the industry could achieve from AI by 2030 and beyond. The real question is: Which banks will seize the opportunity by becoming AI-first?
Banks present many prime opportunities to derive value from artificial intelligence and its most recent incarnation, AI agents—including enhanced customer support, personalization of service and offers, and automation of manual or repetitive tasks. The resulting revenue gains and cost savings would offset the impact of a squeeze on margins in traditional operations from slowing revenue growth and fast-rising costs. (See Exhibit 1.)
While future cost increases could still be offset by revenue growth, near‑term economics reinforce the urgency of taking action. Rates eased in 2025, and savings revenues fell by almost 35% year over year. At the same time, increases in fixed and regulatory‑driven costs (such as compliance, IT, and controls) as well as digital marketing spending continued to outpace total opex, leaving many incumbents with cost‑to‑income ratios of 60% or more, compared to about 35% at well‑run digital banks. Looking ahead, we expect retail bank revenue growth to slow to just 2% to 4% a year through 2029—making a step‑change in productivity imperative.
A small but growing number of banks are delivering value from AI in the form of cost savings, revenue gains, and EBIT increases. Virtual assistants provide customers with financial information and real-time money insights in voice-operated interactive conversations. They are reducing costs and improving conversion rates with no drop-off in customer service ratings. AI agents are handling collections with improved success rates and 30% to 40% lower costs with higher cure rates. Agent-led personalization campaigns featuring dynamic pricing and offers lift ROI and accelerate adoption of fee-based services.
The talk still outpaces the walk, but there are strong reasons to believe that leading retail banks will move toward an AI-first technology foundation and operating model. In addition to the profit potential and margin squeeze cited above is the simple, yet substantial, risk of inaction. Even if only some retail banks successfully embed AI across their operations, these institutions will alter the competitive dynamics of the industry and render non- or slow-movers increasingly irrelevant over time.
More aggressive banks can get a head start from AI agents, which combine predictive and generative AI (GenAI) capabilities to observe, plan, and act autonomously. Unlocking this value, though, requires having the right prerequisites in place, including strong data foundations, scaled AI capabilities, and clear governance.
Hardly mentioned at all in 2024, agents already account for 17% of total AI value in 2025 across all industries, and BCG research shows them reaching 29% by 2028. Agents promise to be the biggest accelerator of value from AI and the basis for the AI-first retail bank. (See Exhibit 2.) Banks that ignore potential advances could be threatened with disintermediation.
Subscribe to our Financial Institutions E-Alert.
The AI-First Retail Bank
AI-first banks will redefine what market-leading banks can do. Deploying AI can remake the cost structure of the institution, but leaders will go further, reshaping existing products and services and inventing new ones. They will become always-on financial partners for their clients, helping people achieve their financial goals and manage their financial lives with autonomous, real-time, data-based suggestions and actions. Implemented with care, agents will collaborate with customers under explicit policy, assurance, and human‑oversight guardrails. In banking, they become a financial operating system embedded across the other ecosystems where customers already live and transact business. Historically, retail banks competed on salesforce strength, efficiency, risk judgment, and the solidity of their balance sheet. Banks that master the transition to AI-first will no longer compete on branch density or balance sheet size, but on the speed, sophistication, and transparency of their algorithms.
Six characteristics will substantially define an AI-first bank:
- Hyper-Personalized Customer Engagement. The branch manager and financial advisor of the future is an AI agent in the customer’s pocket that continuously monitors and engages with the individual’s financial life, anticipating needs before the customer articulates them and making suggestions (or, if authorized, even taking steps) to implement improvements in real time.
- Individual Comprehensive Financial Solutions. Traditional products (such as loans, deposits, and credit and debit cards) will be replaced by adaptive financial solutions that flex in real time according to customer behavior and immediate needs (such as purchasing groceries with a credit card or financing a TV or a car).
- Invisible, Embedded Interfaces. Payments, credit, and savings are embedded seamlessly into apps and ecosystems where customers already live and do business (e-commerce, mobile, social).
- Autonomous Operations. AI agents will supervise and execute end-to-end workflows across service, compliance, risk, and exceptions—driving near-zero marginal cost at scale within clear policy, assurance, and human-oversight guardrails. The focus of a reduced mid- and back-office staff shifts to high-impact analysis and interventions. Ways of working become hybrid by design.
- Real-Time Risk and Capital Allocation. AI agents dynamically steer balance sheets, shifting liquidity, funding, and risk-weighted assets across clients, portfolios, and even geographies in near-real time. This allows banks to maximize return on capital under existing regulatory constraints while simultaneously tailoring solutions for each customer.
- Lean Human Core. Organizations shrink in headcount (and cost) but expand in reach and effectiveness as humans focus on strategy, governance, creativity, and relationships. The AI-first bank may employ fewer people than a mid-sized fintech today yet control more assets than the largest banks of the last century.
Follow the Leaders
The AI-first retail bank and its potential implications may appear far-fetched. It’s certainly true that the reliability of AI models needs to improve, technological infrastructures need to be in place, regulations must catch up, and, most important, consumers and employees have to become comfortable interacting with “smart” machines.
To generate material systemic value from AI, the kind that shows up on income statements and shareholder returns, retail banks must dramatically raise their ambitions and aggressively scale up AI implementation.
Typically, we see companies move through three stages of investment in AI; it is only when they start to scale in the latter two stages that they see P&L impact.
- Deploy. AI automates or augments daily routines throughout large parts of the organization with mainly standard solutions such as document extraction, sanctions screening, and drafting and summarization of text.
- Reshape. AI optimizes processes and functions end-to-end (such as credit process analysis, contract generation, report drafting, and context-specific chatbots).
- Invent. AI enables entirely new business models based on data and algorithms, including personalized financial operating systems, real-time risk pricing, and predictive investment platforms.
The crucial question for most banks is: Will they remain in deploy mode or move forward to invent fast enough to rewrite the rules of banking? Those that fail to move beyond the deploy stage risk being structurally overtaken.
The good news is that leaders in other sectors have developed a proven playbook for getting value from AI that banks can readily adapt. It’s based on five core strategies, although banks need to add one or two of their own owing to the particular attributes of the industry. Download our full report below for details.
AI-first banking is no longer a distant possibility. Banks that actively pursue the path to AI-first in a structured and determined way will secure a lasting advantage as they shape the future of the industry.
