John Cutler · 15 min read · Product Teams
TBM 417: Before You Fire All Your Glue People Because of AI
Read the original on cutlefish.substack.com →
The piece argues that AI adoption at work is risky when we rely on heuristics that can be correct in some contexts but cause lasting damage in others, especially if we replace key 'glue' people without understanding their underlying contributions. It urges a disciplined framework to decide where AI fits and to resist hollowing out essential capabilities, underscoring the importance of knowledge management and clear decision artifacts. It also offers a practical 'Should Do' list to maintain useful practices (docs, research libraries, launch notes) that AI cannot simply replicate.
Key Takeaways
- AI decisions should be context-specific: heuristics can be helpful but can cause structural harm if applied inappropriately, so skepticism and validation are essential.
- Before firing or outsourcing glue roles, map exactly what those people do and what remains when the visible outputs are preserved.
- Preserve underlying processes and artifacts—risk cross-references in docs, curated knowledge libraries, and decision-ready launch notes—to avoid losing capability.
- Develop a practical framework (the 'Should Do' list) to distinguish tasks AI should assist with from tasks that require human-driven knowledge management and collaboration.
Excerpt
(Sometimes I get excited by a topic and write more than once a week. This is a topic I am very passionate about.) The hardest part of figuring out how to use AI at work is that our heuristics are right often enough to feel reliable, and wrong often enough to cause real damage. We develop mental models for where AI fits. Those models work in some contexts. Then we carry them into contexts where they don’t work, and because the logic sounds the same, we don’t notice the difference until it’s too late. People lose their jobs based on…