In 1952, the second season of I Love Lucy began with an episode titled “Job Switching.” After being berated by their husbands for their spending habits, Lucy and Ethel decide to earn their own money by taking a job wrapping chocolates in a candy factory. Things start out smoothly at the candy factor, but the conveyor belt quickly reaches an unmanageable speed, and the women begin shoving chocolates into their mouths, under their hats and into their clothing as they frantically try to keep up with the candy production.
That famous slapstick scene can be seen as an apt metaphor for a very modern problem. As artificial intelligence (AI) enables employees to complete tasks in a fraction of the time that they previously took, managers and executives within the financial industry are learning first-hand about the difficulty of dealing with unmanageable workloads.
AI Productivity: A Double-Edged Sword
AI is enabling teams to produce more credit analyses, compliance reviews, and marketing materials in far less time, but decision-making capacity has not scaled at the same pace. The result is a classic productivity paradox: institutions double the volume of work reaching managers, yet approvals, exceptions, and governance reviews still move at human speed.
Before AI, the time it took to execute tasks and produce deliverables gave managers and executives ample time to review their output and make thoughtful decisions. Today, heightened regulatory scrutiny around AI use means human oversight is more important than ever, but managers cannot simply automate approvals or compress judgment to match AI-driven output. Instead, the bottleneck simply shifts up the chain, increasing the risk of decision fatigue and inconsistent application of policy and procedure.
According to a recent report from McKinsey, AI is already yielding a 20% to 60% increase in workflows for analysis and credit risk, while research by HZ Group finds that this surge in output has created additional pressure on top management, which must oversee governance risks and algorithmic issues. As AI flags more alerts and insights in compliance, and generates more complex credit memos, senior leaders and boards face a growing queue of exceptions and escalations — slowing decisions and, in some cases, affecting the customer experience.
For CFIs, the core challenge is not just adopting AI to increase output but understanding where the bottleneck moves when they do. If underwriting memos or compliance reviews are produced twice as fast but approval timelines remain unchanged, overall throughput has not improved. AI has simply exposed the next operational constraint, often at the management or governance level.
Reframing AI for Management
This constraint is particularly relevant for CFIs, where management layers are thinner and decision rights are more concentrated. As AI accelerates production across credit, compliance, and marketing functions, institutions may find that existing approval structures (not employee productivity) become the primary limiter.
A more measured approach can help align AI-driven gains with institutional capacity:
- Clarify decision rights and escalation thresholds. Routine items can remain decentralized while higher-risk decisions can be elevated.
- Streamline approval layers where appropriate. The workflow can be supported by AI-driven guardrails (e.g., credit thresholds, policy flags) to maintain risk discipline.
- Use AI selectively to help managers triage, summarize, and prioritize incoming work. There are AI features in email and collaboration platforms that can cluster related messages, generate brief thread summaries, and highlight items that require approval or follow‑up.
- Set expectations for quality over quantity. Require initial review or refinement before work is escalated.
For CFIs, the objective is not simply to move faster, but to ensure that gains in productivity translate into smoother decision-making and consistent risk oversight.
AI may be changing the speed of the conveyor belt, but CFIs still control how work moves through the line. By treating AI as a lever for better coordination and clearer decision pathways, not just more output, institutions can surface and address the next operational constraint instead of letting it quietly shift to already stretched managers. With thoughtful governance, calibrated approval structures, and a focus on quality over volume, AI can become less a source of bottlenecks and more a way to reinforce sound credit judgment, compliance discipline, and a more reliable customer experience.
