So far, we have focused heavily on high-level corporate strategy and transformation programs. However, execution happens on the ground, often within specific projects.

In this lesson, we will explore how to scope Cascade’s AI to deliver deep, granular insights for a single project, rather than a broad organizational overview.

The Power of Scoping

The most important concept in this lesson is understanding Input vs. Output. Cascade’s AI analyzes whatever data is included in the report's scope.

To create a Project Report, you don't need a different AI tool; you just need to narrow the scope.

1. Add a Data Table (Discover Widget): Filter this table to show only the specific project you are reporting on (e.g., "Launch Customer Loyalty Program") and its sub-goals.
2. The AI's Reaction: Because the report now contains only data relevant to this one project (tasks, KPIs, updates), the AI automatically ignores the rest of the business and focuses its narrative entirely on this project's health.

Anatomy of a Perfect Project Report

For a project-level view, we recommend a structure that moves from Context to Data to Insight.

1. The Header (Static Context)

Start with standard text widgets to set the scene. This replaces the title slide of your old PowerPoint deck.

• Scope Statement: What is this project delivering?
• Core Team: Who is responsible?

2. The Scoreboard (KPIs)

Immediately below the header, place your key metrics (e.g., Budget Burn, ROI, % Complete).

• Tip: If you have integrations with Power BI or Tableau, display those live charts here.

3. The AI Insight Layer (Side-by-Side)

This is where you replace manual status writing. A powerful configuration is placing two AI widgets side-by-side:

Widget A: The Status CheckWidget B: The Fix-It PlanConfig: Executive Summary (Default)Config: Custom PromptRole: This acts as the project manager's update. Because the report is scoped to one project, it mixes quant data (deadlines) with qual data (recent updates) to summarize current progress.Role: This acts as a troubleshooter. Use a custom prompt to identify specific remedial actions for this specific project.

The 'Fix-It' Custom Prompt

In the transcript example, the right-hand widget wasn't a standard configuration. It was a Custom Prompt designed to turn issues into tasks.

Try this Prompt:

"Review the data for this project. Identify the top 2 critical issues currently blocking progress. For each issue, suggest 3 specific, actionable steps we can take immediately to get the project back on track."

The Output:

Instead of a generic summary, the AI will list:

1. Issue 1: (e.g., Vendor delay)
   • Action: Schedule urgent review.
   • Action: Source backup supplier.
2. Issue 2: (e.g., Budget overage)
   • Action: Review non-essential scope.

From PowerPoint to Dynamic Reporting

The rule of thumb for Cascade Reporting is simple: If you used to put it on a PowerPoint slide or an Excel tab, you can replicate it here.

• Want a list of completed tasks grouped by owner? Prompt for it.
• Want a summary of risks specific to the engineering team? Prompt for it.

As long as the data (the tasks, the updates, the risks) exists in Cascade, the AI can structure it however you need.

Summary

• Scope Control: Use filtered data tables to focus the AI's attention on a single project.
• Layering: Build your report logically: Context -> Hard Numbers -> AI Analysis.
• Custom Prompts: Use custom prompts to ask specific questions about the project, such as "How do we get back on track?" rather than just "How are we doing?"