Personal agents vs. company agents: who owns the context?
The difference is not how smart the model is. It is who the system serves, what it can remember, and who is accountable.
A point of view.
A way to check it.
Research and practical perspectives on making AI useful inside a company.
7 insights
3 min read
Your team uses AI. Has the way your company works actually changed?
Individual usage is a starting point. Operational change needs a workflow, an owner, and a before-and-after comparison.
What belongs in a company brain—and what doesn’t?
A research note on useful context, maintained sources, and the difference between remembering and knowing.
Read, draft, act: deciding what AI can do without asking.
A practical way to discuss permissions before a useful assistant becomes an unpredictable operator.
What Uber’s Genie teaches us about a company brain.
Less waiting for answers. Fewer repeated questions. A practical lesson in making company knowledge useful.
Same models. Different depth of adoption.
An operator’s reading of OpenAI’s Enterprise Signals: the usage gap, the skills gap, and the work behind both.
The usage gap widened
Frontier / typical output tokens per active user · multiple
Jan 20262.6×
Jun 20268.3×
The environment changes what people delegate.
What Anthropic’s Cadences report suggests about tools, delegation and learning—and what it cannot prove.
The delegation gap survives a same-model comparison
Claude Code minus chat / Cowork · points on a 1–5 autonomy scale
Overall0.37 pts
Sonnet only0.26 pts
No matching notes yet.
Try a broader term or clear the filters to see the full library.
Research notes distinguish the source’s findings from our interpretation.