daBongo LMS AI Training Courses

Lesson Content

The format problem.

A Gemini response that contains excellent information in the wrong format requires rework before it is usable. A comparative analysis written as flowing prose needs to be manually converted to a table. A numbered list that should have been a framework requires restructuring. Format specification is not a minor formatting concern – it is the difference between immediately actionable output and output that needs hours of reformatting.

Six high-value output formats.

1. Tables: Best for comparisons, option analysis, feature lists, and structured data. Request: "Format this as a table with columns: [column 1], [column 2], [column 3]."

2. Numbered lists: Best for steps, ranked options, or ordered processes. Request: "Give me this as a numbered list. Each item should be one sentence."

3. Bullet lists: Best for unordered collections of items, features, or considerations. Request: "Format as bullets – each bullet under 15 words."

4. Hierarchical outlines: Best for document structures, presentation frameworks, and organized content. Request: "Format as an outline with main headers and sub-points under each."

5. JSON or structured data: Best for technical users needing data in a parseable format. Request: "Format the output as JSON with these fields: [field names]."

6. Section-and-paragraph structure: Best for formal documents, reports, and communications with specific sections. Request: "Format with section headers [list headers] and 2-3 paragraphs under each."

Requesting parallel structure.

When you need multiple items in the same format – multiple options, multiple analyses, multiple recommendations – ask for parallel structure explicitly:

"For each of the three options, give me: Option Name, One-Sentence Description, Top Benefit, Top Risk, and Estimated Timeline. Use the same structure for all three so I can compare them side by side."

Parallel structure is the fastest way to create a true comparison – and it almost never happens by default without explicitly requesting it.

Matching format to use case.

Before specifying format, ask: where does this output go next? If it goes into a presentation, request a slide outline. If it goes into a spreadsheet, request a table with specific headers. If it goes into a client email, request prose with clear headers. The format should be specified based on its destination, not on what seems easiest to request.

Combining format with other components.

Format requests are most powerful when combined with the full five-component structure:

"I am preparing a competitive analysis for our leadership team [Context]. Compare our three main competitors on these dimensions: pricing model, target customer, key differentiator, and biggest weakness [Task]. Act as a market analyst who is being honest about where we are behind [Role]. Do not speculate – flag anything you are uncertain about [Constraint]. Format as a table with competitor names as columns and dimensions as rows, plus a brief paragraph summary of the most important takeaway [Format]."