Step 2: Determine the data you need

Step 2: Determine the data you need

Why context matters

You must identify the data required to make the model context-aware. Without enough context, the model will produce vague or inaccurate outputs.

Signs of insufficient context

• Generic phrases or ambiguous nouns (e.g., “Systems,” “Automations,” “Insights”)
• Inaccurate descriptions or incorrect personification of concepts
• Ability to discuss a topic in general but not address specifics
• Tendency to “beat around the bush” rather than deliver concrete details

Breaking down output components

For each component of your final output, list the specific data the model needs:

• Tags
  • Provide a list of all possible tags for the title
• SERP research and example titles
  • Use these as input to inform tone and style
• Header image
  • Supply an image description to generate the visual

Identifying shared data

Some data elements may serve multiple components. As you decompose your output, note which pieces of information are reused across tags, tone/style inputs, images, or other sections.

In the next video, we’ll walk through the next step which is generating the content after we’ve decomposed every part of the output.