Two years after generative AI captured Wall Street’s attention and investment dollars, banking executives are demanding answers about return on investment. According to Larry Lerner, a partner in McKinsey’s banking practice, this has become “the $20 billion question” facing the financial services industry.
While banks assembled teams of technologists to experiment with generative AI and run proofs of concept, many institutions now find themselves stuck in what Lerner calls “POC purgatory"—trapped in endless experimentation without seeing tangible results. According to an October report from Evident AI, only six out of 50 banks disclosed actual dollar-level cost savings or revenue increases from their AI investments.
However, a handful of frontrunner firms are beginning to see tangible returns through current cost savings, future cost avoidance, and incremental revenue generation. McKinsey’s fresh research identifies several key factors separating winners from laggards in the AI investment race.
Critical success factors include treating AI as a business opportunity rather than a technology problem, with business leaders—not just tech teams—held accountable for results. Banks that concentrate their efforts on a handful of high-impact use cases instead of spreading resources across dozens of small experiments are seeing faster paths to value.
Measurable ROI is emerging in specific areas: AI-powered call center copilots and analytics platforms for customer targeting are delivering quantifiable results. One large bank projects a 10% revenue increase from a new analytics platform, while fintech company Klarna estimates its OpenAI-powered call center agent will generate $40 million in profit this year, doing the work of 700 full-time agents.
Lerner notes that banks are beginning to modify forward-looking hiring plans, particularly in contact centers, as AI enables increased self-service and faster resolution times. This “cost avoidance is absolutely measurable,” he emphasized.
Key recommendations include building reusable AI tools that can be deployed multiple times across the organization, accelerating development while satisfying risk, security, and compliance requirements. However, the ultimate challenge remains adoption—getting workers and customers to embrace new AI-powered workflows, an obstacle that has plagued previous technology cycles in banking.
Key Quotes
That is the $20 billion question
Larry Lerner, McKinsey partner, describing the urgent pressure bank leaders face to demonstrate returns on their massive AI investments after two years of experimentation and spending.
Instead of having 60 use cases across 15 different business lines and functions, narrow down to three areas where you want to go deep
Lerner explaining why concentrated AI efforts in a few high-impact areas deliver faster value than spreading resources across numerous small experiments—a key lesson for banks stuck in POC purgatory.
The value of what you’re doing depends on how you’re going to repurpose your time, and that’s really hard to do. Because it’s an indirect sort of lever, it’s very difficult to actually measure and get people to agree that there’s value.
Lerner addressing the challenge of measuring productivity gains from AI, explaining why time-saving benefits often fail to translate into demonstrable bottom-line impact that satisfies executives.
Most companies have done a pretty bad job of getting adoption to the level that’s going to yield the results that they want to yield
Lerner identifying user adoption as the critical bottleneck preventing AI investments from delivering expected returns, echoing challenges from previous technology transformation cycles in banking.
Our Take
The banking sector’s AI reckoning reveals a universal truth about enterprise technology adoption: innovation without execution yields no value. McKinsey’s findings expose the dangerous middle ground where organizations invest heavily but lack the strategic discipline to convert experiments into scalable solutions.
The “POC purgatory” phenomenon reflects deeper organizational challenges—siloed decision-making, unclear accountability, and the persistent tendency to treat transformative technologies as IT projects rather than business imperatives. The 12% success rate among banks should alarm AI vendors and enterprise buyers alike.
What’s particularly telling is the divergence between productivity-focused AI tools and revenue-generating applications. While time-saving copilots struggle to demonstrate ROI, customer-facing AI like Klarna’s call center agent delivers measurable millions. This suggests the first wave of AI winners will be those solving problems with direct financial metrics—customer acquisition, retention, and service cost reduction—rather than abstract productivity gains. The banking sector’s experience will likely accelerate a broader industry shift toward accountable, measurable AI implementations.
Why This Matters
This story represents a critical inflection point for the AI industry as early adopters transition from experimentation to accountability. With billions invested in generative AI over the past two years, the financial services sector’s experience serves as a bellwether for enterprise AI adoption across industries.
The emergence of “POC purgatory” highlights a fundamental challenge facing organizations worldwide: translating AI’s theoretical potential into measurable business value. The fact that only 12% of banks can demonstrate concrete financial returns underscores the gap between AI hype and reality.
For the broader AI ecosystem, this signals a maturation phase where vendors and technology providers must deliver solutions with clear ROI rather than promising future capabilities. The success stories—like Klarna’s $40 million profit impact—demonstrate that AI can deliver substantial returns when properly implemented in high-value use cases.
This shift toward accountability will likely reshape AI investment priorities, favoring measurable applications like customer service automation and revenue-generating analytics over productivity tools with indirect benefits. The banking sector’s lessons will inform AI strategy across healthcare, retail, manufacturing, and other industries facing similar pressure to justify their AI spending.
Related Stories
- Wall Street Asks Big Tech: Will AI Ever Make Money?
- The AI Hype Cycle: Reality Check and Future Expectations
- How Companies Can Use AI to Meet Their Operational and Financial Goals
- Tech Workers Are the Real Winners in the AI Talent War, With Pay Set to Soar by 2024
Source: https://www.businessinsider.com/how-wall-street-banks-can-find-roi-ai-investments-mckinsey-2024-12