Converting Content Formats

Identify source format, target format, and purpose, then restructure — don't just reformat:

Source: Table (sales data by region/quarter)
Target: Text (executive summary)
Action: Extract trends and key numbers → write narrative with topic sentences,
        not a row-by-row transcription of the table

Progress:

• Step 1: Identify source format and target format
• Step 2: Identify the audience and purpose of the target format
• Step 3: Extract the core content units (facts, relationships, arguments)
• Step 4: Re-derive structure appropriate to target format (not a literal transcription)
• Step 5: Verify no data loss on critical facts (numbers, names, dates)
• Step 6: Verify target format conventions are followed (slide density, table headers, paragraph flow)

Step-by-step by conversion type

Table → Text:

1. Identify what the table is about (the implicit thesis)
2. Group rows/columns into logical clusters (trends, categories, outliers)
3. Lead with the most important insight, not row 1
4. Convert numbers into comparative language ("grew 30%" vs restating every cell)
5. Keep precise figures for key data points; summarize the rest

Text → Table:

1. Scan for repeating attributes across entities (these become columns)
2. Identify the entities themselves (these become rows)
3. Define column headers that are consistent in unit/type
4. Extract one fact per cell — split compound statements
5. Flag text that doesn't fit tabular structure (keep as footnote/caption)

Text/Table → Slides:

1. Find the narrative arc or logical sequence (this becomes slide order)
2. One idea per slide — split dense paragraphs across multiple slides
3. Convert sentences into fragments/bullets (verb-first, no articles when possible)
4. Convert tables into simplified visuals: show only the 3-5 rows/cols that matter
5. Add a title per slide that states the takeaway, not just the topic
6. Reserve detailed data for an appendix slide or speaker notes

Slides → Text/Table:

1. Reconstruct full sentences from bullet fragments (fill in connecting logic)
2. Merge related slides into unified sections
3. Recover context lost from speaker notes if available
4. Rebuild any simplified slide data into full tables if precision is needed

Example 1: Input (Table):

Output (Text): "Regional sales diverged sharply over the year. The North region led for most quarters but dipped in Q3 before rebounding strongly to 160 in Q4. The South region showed steadier growth, nearly doubling from 90 in Q1 to 140 in Q4 — including a standout Q3 that briefly outpaced the North. Overall, both regions ended the year at their highest levels, suggesting broad-based momentum."

Example 2: Input (Text): "Our three main competitors differ in pricing and focus. Acme charges $50/month and targets enterprise clients. Beta charges $20/month and focuses on small businesses. Gamma is free but monetizes through ads, targeting individual consumers."

Output (Table):

Example 3: Input (Text, dense paragraph): "The project faced three major risks: budget overruns due to scope creep, timeline delays from vendor dependencies, and quality issues stemming from insufficient testing cycles. Each risk was mitigated through specific controls — a change control board for scope, dual-sourcing for vendors, and an expanded QA phase."

Output (Slides):

Slide 1: "Three Key Project Risks"
- Budget overruns (scope creep)
- Timeline delays (vendor dependencies)
- Quality issues (insufficient testing)

Slide 2: "Risk Mitigation Strategy"
- Scope → Change control board
- Vendors → Dual-sourcing
- Quality → Expanded QA phase