Explanation
Kno and your eval framework
Why Kno does not compete with eval frameworks, it consumes them, and where the division of labor sits.
Your eval framework answers: does version B perform better than version A? That answer is the foundation Kno stands on.
Kno answers the next question: which of these 500 data assets caused the improvement, how confident are we, what did it cost, and where should each one live?
Eval framework: "Does version B perform better than version A?"
Kno: "Which of these assets caused the improvement,
how confident are we, what did it cost, and where
should each one live?"
The division of labor
| Capability | Eval framework | Kno |
|---|---|---|
| Model performance | ✓ | uses it |
| Prompt performance | ✓ | uses it |
| Asset-level impact | , | ✓ |
| Marginal data value | , | ✓ |
| Data cost | , | ✓ |
| Destination selection | , | ✓ |
| Portfolio validation | , | ✓ |
Kno re-scores your agent with each candidate asset, holding the model, prompt, and tools fixed, and reads the result through your evals. The two tools compose; they do not overlap.