Skip to content
Culsma PAWS

Ask Questions About Project Data

Use this chapter for counts, comparisons, trends, and missing-data questions about an existing Lab Project.

These questions are read-only. Start by naming the project and the measure you want. Include a time range, grouping, or filter when it affects the meaning of the result.

Example questions

Section titled “Example questions”

Count samples by month for this project.

Compare the number of samples across sample types.

Which supported fields are most often missing in this project?

For Project Alpha, compare PBMC and plasma sample counts by collection month in 2026.

A simple workflow

Section titled “A simple workflow”
  1. Ask the AI agent to confirm the active workspace, source Lab Project, and available fields.
  2. Clarify the metric, date field, grouping, and exclusions.
  3. Review the effective filters and matching sample population.
  4. Review the result together with its assumptions and completeness notes.

If the request cannot be answered reliably from the available Lab Project records, the AI agent should say so. Do not treat an estimated, truncated, or partial answer as a complete project result.

Read the answer carefully

Section titled “Read the answer carefully”

Confirm the selected project, grouping, time range, assumptions, and incomplete data before interpreting a result.

A useful answer should include the effective filters, number of records examined, missing values, and whether any results were skipped or truncated.