After this lesson you can add a Yes / No, Score or Pick one column that asks Jev a question about each row, and filter on the answers.
1:31 video with captions, written steps and transcript. Checked against Duckrow on October 3, 2026.
AI Columns and Jev · Lesson 5
90 seconds
Narration: preview voiceA private equity team maps add-on targets for an HVAC platform. The companies are example data.
Click Add column and choose Yes / No, Score or Pick one. Write the question Jev asks about each row; for Score, list the levels lowest first, and for Pick one, the options. Each row's answer is a plain value, so you can filter on it. It costs 1 credit an answer on Duckrow's Jev, charged only when Jev answers, and nothing with your own Jev key.
About five minutes, in your own workspace.
You're done when your table has a Yes / No column you trust enough to filter on.
Three quick questions. Nothing is saved.
Some questions just need a yes or a no. Jev columns ask one question about every row, and give you an answer you can filter.
Click add column. Under Enrich, there are three Jev columns: yes or no, score, and pick one, each one credit a row. Choose yes or no.
Type the question Jev asks about each row, using its other columns: does it do commercial HVAC work? The question becomes the column name. Run on new rows answers new imports as they arrive. Then add the column.
Run it from the column header. Each cell shows running, then a plain yes or no. That means a filter, equal to yes, keeps exactly the commercial firms.
Score works the same way, with levels you list lowest first, such as low, medium and high. Pick one chooses from your options, one per line.
Each answer is one credit on Duckrow's Jev, charged only when Jev answers. With your own Jev key, answers are free.
To recap. Yes or no, score and pick one ask Jev one question per row. The answers are plain values you can filter. One credit an answer, free on your own key. Next, spend nothing on rows Jev rules out.