Core Workflows

How to Ask Better Questions in DashboardGenius

Prompt frameworks and examples for clearer, faster, and more actionable manufacturing insights.

How to Ask Better Questions in DashboardGenius

Answer quality improves most when you increase scope clarity.

If you do not know the right source, metric, timeframe, or grain yet, start in Explore. If you already know those pieces, use Analyze directly.

Use the 4-Part Prompt Structure

Include:

  • Metric: what to measure
  • Time: exact period
  • Scope: line, area, product, plant
  • Output: trend, rank, compare, summarize

Template:

"Show [metric] for [time] by [scope], and [output]."

Strong Prompt Examples

  • "Compare OEE by line for the last 14 days and rank highest to lowest."
  • "Show top downtime categories on Line 2 for night shift last week."
  • "Trend quality rejects by SKU family month to date and call out the biggest change."
  • "Compare this week vs last week for throughput by area."

Better Follow-Ups (Instead of Rewriting Everything)

After the first answer, refine with one focused follow-up:

  • "Break this down by shift."
  • "Filter to Plant 3 only."
  • "Compare to prior period."
  • "Show top 5 only."

This preserves context and gets to useful analysis faster.

When to Use a Widget Instead

If the question follows a repeatable pattern such as capacity planning, downtime pareto review, yield analysis, shift summaries, or schedule risk checks, start with a widget.

Widgets are useful when you want stronger framing without having to write the full request yourself.

When to Use Explore Instead

Use Explore when the prompt is not ready for a final answer yet.

Good Explore starts include:

  • "Help me scope the best way to compare demand and inventory for next week."
  • "Which connected source should answer customer-margin questions?"
  • "I want to analyze a product family, but I am not sure what it is called in the data."

After Explore creates an Analysis Brief, review the suggested source, timeframe, and grain before running the full analysis.

Common Prompt Mistakes

Avoid:

  • Missing or vague time ranges ("recently")
  • Multiple unrelated requests in one message
  • No business scope (line, area, product, plant)

If results look broad, tighten one variable at a time.

Reusable Prompt Patterns by Use Case

Daily operations check

"Summarize yesterday's production performance by line with top downtime drivers."

Quality review

"Show reject trends for the last 30 days by product family and highlight recurring failure themes."

Leadership snapshot

"Give a weekly executive summary: OEE, downtime, quality rejects, and top improvement opportunities."

Team Standardization Tip

Define 5 to 10 canonical prompts for regular meetings (shift handoff, weekly quality, leadership review). This improves consistency across teams and reduces rework.

Next Guide

Continue with Explore Before Analysis.