Marketing Attribution Funnel Generator

Problem

Marketing attribution reports often describe exposure, click, registration, and conversion as free text. Manually extracting touchpoints, weights, conversion rates, and inferred values can be inconsistent, especially when some numbers are missing or need to be labeled as inferred. This skill turns a marketing report into reusable attribution artifacts: a funnel image, structured data, and the prompt used for image generation.

How It Works and Limits

• Intent parsing: identifies the funnel levels and key metrics in the report.
• Extraction and inference: extracts real values first; missing fields can be inferred from context and marked with is_inferred.
• Async generation: submits a task via execute_task.py, receives a task_id, then polls with query_task.py; each poll response should be surfaced immediately.
• Deliverables: returns funnel_image_url, structured_funnel_data, and optimized_prompt, with optional HTML preview and aspect ratios such as 16:9 or 9:16.

Limits: report_text is required; richer input yields a more precise funnel. Real calls require a ShowAPI appKey; without one, the result is only a simulated preview. The flow is strictly asynchronous and is not intended for highly custom interactive charting.