Generative AI Document and PDF Prompts

How it works

Document workflows often include an extraction step before language generation. Reading order, footnotes, scanned text and table headings can be lost. Ask the tool to identify the evidence it used and report unavailable content rather than pretend the whole file was read. For a real PDF, inspect extracted rows against the original pages before using numbers in another system.

How to use this lesson

This is an authored prompt exercise with a human-written acceptance guide. It is not a captured provider response or a promise that a model will produce the target. No provider account, API, model weights or live inference was used for the course execution review. If you try it in an external service, use an authorised environment, check its actual access and data settings, and record the product/model and run date.

Bhylu does not run a generative model or send this prompt to an AI service. Copy and download preserve the text only. External tools may require accounts, paid access or compatible inputs; this course does not promise availability or prices.

Prompt exercise

Download the prompt

Using this extracted table only, report quantities by kit.
Kit | Sheets
L2 | 4
L3 | 6
Do not claim to have read an attached PDF; no PDF is attached.
Keep kit identifiers and counts exact.

Acceptance guide — not a model transcript

Reference: L2 has 4 sheets and L3 has 6.

The fixture is intentionally explicit about its source type. A real file workflow should record extraction status, page coverage and any unreadable regions. Never let the model use a missing row as evidence of zero unless the business rules actually define that meaning.