How do I set up an AI Interpretation for a cohort report?

A cohort AI Interpretation generates written feedback from aggregated cohort data, such as a 360, group, team, or iterative report. The setup mirrors an individual AI Interpretation, with one key difference: it must live in a section set to the Cohort Report type.

How to set one up

1. Add a cohort report section. In the Assessment Builder, add a new Section and save it. Then open Edit Section > Options, set Section Type to Cohort Report, and save. This section is hidden from respondents.

2. Add the AI Interpretation. Add a subsection inside that section, then click Add AI Interpretation. The interpretation now sits within the cohort report section, so it runs against aggregated cohort data.
3. Write the prompt. Enter the instructions and use Add Merge String to feed in the cohort's data. For example:

You are an organizational development consultant summarizing a team
leadership assessment for a cohort.

Here are the aggregated cohort results:
{Tables[SectionNo=ALL]}

Write a concise summary (3 to 4 short paragraphs) that:
- identifies the cohort's strongest and weakest sections
- highlights notable patterns across the group
- suggests two or three team-level development priorities

Treat the data as aggregated and do not single out individuals.
Use a professional, constructive tone and bold the section names.

4. Test it. It is best to have an existing cohort to test against. Select that cohort from the dropdown to preview the result using real aggregated data, then adjust the prompt until the output reads the way you want.
5. Add the merge string to your report. Same as the individual case: insert the Answer Text merge string for the interpretation's position, for example {ResponseAnswer.AnswerText[S5 SS1 Q10]}.

How do I view a cohort's AI Interpretation result?

To see what the AI generated for a cohort, and to review it before the cohort report is sent:

1. Go to the Cohort and navigate to the cohort you want.
2. Click View.
3. Open the Interpretations tab. This displays all generated results for that cohort.

Click Edit to review, adjust, and lock a result. Do this before the cohort's requested completion date, as the locked result is what gets generated into the cohort report.

For the full setup, see Creating AI Interpretations for Reports. For prompt-writing best practices, see Writing Effective AI Interpretation Prompts.