Table of contents
In my last post, I wrote about swamping, one of the three reasons every organization needs an alignment agent. Everyone can generate far more content with AI, but we can only read it slightly faster.
The second reason is another cost: re-asking.
The good news about AI is that everyone can do their own research to understand what is going on.
The bad news is that… everyone can do their own research to understand what is going on.
That costs you in three ways:
- An opportunity cost: The same question could be answered once.
- An accuracy cost: Not everyone asks the question the same way, or with the same context.
- A real cost: Look at your AI spend.
Glean measured this in its Work AI Index 2026: workers spend 6.4 hours a week “botsitting” AI, feeding it context and checking its output.
I can’t tell you how many times I’ve had the conversation I recently had with an engineering leader. He asked, “Are we measuring the right AI adoption metric?” The metrics lived in a few different places. Different teams defined the same metric differently. There was no headline number. That is what re-asking looks like at scale: many teams, many answers, not one correct one.
Swamping and re-asking are two of the three reasons to build or buy an alignment agent.
