Insights are the structured data points that Vocalis extracts from each conversation. Instead of just giving you a raw transcript, the AI Actor pulls out the specific information you care about — scores, key answers, and assessments.
You configure Insights in Step 3 of the AI Actor setup wizard.
Adding an insight
Click Add Insight and fill in three fields:
| Field | What it means |
|---|
| Name | A label for the data point (e.g., “Years of experience”, “Overall recommendation”). |
| Type | How the data is displayed — Text, Score, Bar Chart, Line Chart, or Pie Chart. |
| Purpose | A description that helps the AI extract the right value from the conversation. |
Insight types
| Type | Best for | Example |
|---|
| Text | Open-ended answers, summaries | ”Background summary”, “Salary expectations” |
| Score | Numeric ratings, years, counts | ”Communication rating (1-5)”, “Years of experience” |
| Bar Chart | Comparing categories | ”Skills breakdown by area” |
| Line Chart | Trends or progression | ”Experience level over time” |
| Pie Chart | Distribution or composition | ”Technology stack breakdown” |
Writing good purpose descriptions
The purpose description tells the AI what to look for. The more specific you are, the more accurate the extraction.
Name: Salary expectations
Purpose: The candidate's stated compensation range, including
base salary and any mention of equity or bonuses.
Name: Salary
Purpose: Salary info.
Example insights
Here is a set of insights for a technical interview AI Actor:
| Insight name | Type | Purpose |
|---|
| Background summary | Text | 2-3 sentence summary of the candidate’s experience |
| Primary technologies | Text | Main languages, frameworks, and tools they use |
| Years of experience | Score | Total years in the relevant field |
| Salary expectations | Text | Stated compensation range including base and equity |
| Communication rating | Score | 1-5 rating of how clearly they explain technical concepts |
| Overall recommendation | Text | Strong Yes / Yes / Maybe / No with brief justification |
Start with 6-10 insights. Too few and you miss useful data. Too many and the AI spreads its attention thin. Focus on the data points your team actually uses to make decisions.
Next steps
