Generate_Insights: Add AI insights detection to core functions

[moralec]

Baseline
Our core functions (e.g. collab_sum(), afterhours_trend()) currently generate visuals and summary tables. Actual data interpretation still depends entirely on the analyst. We can improve core functions so they can automatically detect insights.

Idea
Add additional return option to core functions to return a curated bullet list of insights.

For example, collab_sum(hrvar="Organization", Return="text") can return:

- The average employee in this company spends 12 hours in meetings and 8 on emails.
- "Finance" employees spent the most time in meetings (18 hours a week, 1.5x more than the average)
- "Sales" employees primarily rely in email as their preferred collaboration tool (12h a week - 4 more than the population average)
- "Operations" employees have the highest collaboration load: their collaboration time is 30h (10 hours more than the population average)

Implementation

• Augment core function so they can identify some key insights.
• Insights are to be extracted from the key summary table that the function already calculates.
• To keep core functions clean and avoid repetition, we will add supporting functions (create_insights_sum()) to do the heavy lifting. One function will exist per summary data table type (Summary, Dist, Trend)
• Insights will include population averages and outlier detection.
• Conditionals (Top and bottom thresholds) will be used in the function to select what insights are worth including/excluding.
• We will follow the same return structure as the validation functions. A "return=message" option across all functions will print the insights in the console. A "return=text" option across all functions will return insights as text, to be used in reports (HTML and PPT).