A "meat proxy" forwards AI-generated content without adding critical thinking. To provide value, leaders must engage with the output, ensuring it reflects their judgment, rather than just acting as a human pass-through for the technology.

AI makes it easy to generate lengthy documents with little thought. This creates a "processing tax" where the burden of review is pushed onto others, especially executives. Leaders should implement a "slop policy" that treats this behavior as disrespectful of colleagues' time.

Relying on AI without applying critical thinking produces "work slop"—outputs that look polished on the surface but lack genuine depth or substance. This can be dangerously misleading and devalues the quality of work by giving a false sense of security.

Junior team members can easily rely on AI for answers, producing generic "AI slop." Effective managers must actively coach them to critique and augment AI outputs with their own thoughts and opinions, reinforcing that they were hired for their individual judgment.

Senior leaders find AI accelerates work but encourages low-quality, uncritical outputs—a phenomenon called 'AI sloth'. To maintain standards, some build AI personas embodying their own perspective, which teams use to vet work before submission, counteracting the deluge of 'junk'.

To counter the flood of low-effort, AI-generated content, companies like Clay are creating official policies. These guidelines don't ban AI but enforce accountability, ensuring employees use it as a tool for better thinking—not a shortcut that disrespects readers' time by producing verbose, low-value documents.

To combat a flood of low-quality, AI-generated documents, some leaders are creating custom AI agents that embody their personal review criteria. Product managers are required to filter their work through this 'leader persona' skill, which forces them to address key strategic questions and embed critical thinking before submitting for review.

The initial hurdle of getting teams to use AI is over. The more difficult problem is ensuring they use it correctly to produce high-quality work, not just 'slop'—low-effort content generated by outsourcing taste and judgment to models like Claude.

'AI Slop' flourishes when leaders don't explicitly define what 'good' looks like. This failure, combined with decentralized tool usage and a lack of a central source of truth, allows low-quality, AI-generated work to become the default standard within an organization.