As artificial intelligence becomes increasingly embedded in corporate operations, one of Australia’s largest financial institutions is sounding the alarm about a troubling trend: skyrocketing AI costs paired with deteriorating output quality that executives are calling “work slop.”
What Happened
Matt Comyn, chief executive of the Commonwealth Bank of Australia, delivered a pointed critique this week during a corporate leadership speech, identifying two critical problems threatening the ROI of enterprise AI investments. The executive highlighted how token-based billing models—where companies pay per unit of AI processing—are creating unexpectedly massive bills as task complexity increases. Simultaneously, the quality of AI-generated content flowing through corporate workflows has declined noticeably, leading Comyn to characterize much of this output as “work slop.”
The comments represent a significant moment of candor from a major financial institution’s leadership, breaking through the hype surrounding AI adoption to address real operational challenges enterprises face daily. Unlike the breathless startup narratives dominating tech media, Comyn’s observations reflect the ground-level reality many Fortune 500 companies are experiencing after months of aggressive AI implementation.
Key Points
The CBA chief identified several interconnected issues. First, the cost structure of modern large language models creates perverse incentives for overuse, with expenses scaling unpredictably as organizations attempt more complex tasks. Second, the quality-versus-cost tradeoff has become increasingly unfavorable, with companies discovering that cheaper models produce inadequate results, while premium models drain budgets faster than anticipated.
Comyn’s “work slop” characterization cuts through corporate jargon to describe a genuine problem: AI outputs that technically function but lack the nuance, accuracy, and business value that organizations expected. This isn’t about AI hallucinations or occasional errors—it’s about systemic degradation in usability when scaling across multiple business units.
What This Means
The Commonwealth Bank’s CEO comments signal a broader reckoning within corporate America. The initial AI gold rush—where enterprises invested heavily in implementation without fully understanding costs and quality implications—is yielding to a more sobering assessment. Organizations must now grapple with the economic reality that AI adoption isn’t the cost-cutting measure many promised.
This moment matters for several constituencies: CFOs will face increased scrutiny on AI spending justification; enterprise software vendors must address cost efficiency and output quality; and policymakers should note that token-based billing models may not serve large-scale corporate deployment. For the broader tech industry, Comyn’s comments represent a necessary corrective to overhyped narratives about AI’s immediate transformative power.
The conversation is shifting from “if” enterprises should adopt AI to “how” they can do so economically and effectively.
