A spate of surveys paints a sobering picture for enterprise AI: employees and consumers don’t trust AI search results, many pilots fail, and most companies report little bottom-line impact. The latest indictment is “workslop,” a term from a Harvard Business Review study describing AI-generated content that looks polished but doesn’t move work forward.
The column argues the problem isn’t the tools as much as management. Employers often chase hype without standards, training, policies, or ownership. To make AI pay, firms should designate accountable leaders, standardize assistants, set clear use policies, train staff in prompting and review, and track concrete productivity metrics. AI is a tool, not a strategy; when it produces junk, the buck stops with the boss.
Related articles:
NIST AI Risk Management Framework
Microsoft Copilot Adoption Guide
AI Incident Database
OECD AI Principles
