How Bazaarvoice uses Arize AI
10 minutes
“With LLM-based evals, Kratz says the same volume can be evaluated in about 10 minutes.”
As published on arize.com. Checked by usedby on Oct 16, 2026.
What happened
Bazaarvoice uses LLM-based evals to classify the outputs of its AI question-and-answer system as correct or incorrect, and to check whether two answers say the same thing. They validate the evaluators against a human-labeled set and then run them on much larger datasets instead of labeling manually.
Summary written by usedby from the source page, in English. The figures are those of Arize AI and Bazaarvoice, not ours.
- a day and a half“The process took about a day and a half and came with the associated cost of manual labeling.”
For one experiment, Bazaarvoice sent 1,000 examples for human review. The process took about a day and a half and came with the associated cost of manual labeling. With LLM-based evals, Kratz says the same volume can be evaluated in about 10 minutes.
It’s saving us a lot of time on my team.
What the story claims, and what we checked
What we compared with the page.
- The figure: 10 minutesCheckedPrinted word for word on the page, near the name of Bazaarvoice.
- The passage quoted aboveCheckedCopied word for word from the page, near the name of Bazaarvoice.
- Bazaarvoice uses Arize AICheckedConfirmed line.
- The result itselfNot checkedWe quote it; we did not measure it.
More stories
Same company, same tool or same industry.
