AI-assisted practical guide. Examples are hypothetical; these are proposed editorial methods, not reported research results.
Using artificial intelligence to audit your technical explanations against a rubric helps identify gaps in clarity or missing requirements before a final submission. By treating the AI as a critical reviewer rather than a writer, you can isolate specific areas where your prose fails to meet established standards. This process transforms a subjective rubric into a structured gap analysis, with the aim of checking that every required element is present and sufficiently detailed.
Structuring the Comparison Prompt
To get a precise audit, you must provide the AI with three distinct components: the original rubric, the explanation you wrote, and a strict instruction to cite evidence. Suggest that the AI creates a mapping where it lists each rubric criterion and then quotes the exact sentence from your text that satisfies it. If no such sentence exists, the AI should explicitly state that the requirement is unmet. This can help prevent the AI from hallucinating a general sense of quality and forces it to rely on the actual text provided.
Hypothetical example
Imagine you are writing a guide on server migration and your rubric requires a clear warning about data backup. Your prompt might be: Analyze the attached explanation using the provided rubric. For every criterion, quote the specific evidence from my text. If evidence is missing, mark it as a gap.
Input: Rubric: Must include a backup warning. Text: First, shut down the server. Then, move the files to the new host. Finally, restart the system.
Output: Criterion: Backup Warning. Evidence: Not found. Gap identified: The text fails to mention backing up data before shutdown.
Verifying the Audit Results
The most difficult case occurs when an AI claims a criterion is met based on a vague implication rather than a direct statement. To handle this, you should challenge the AI to explain why a specific quote satisfies a complex rubric point. Once the AI provides its analysis, perform a manual human review by highlighting the quoted evidence in your original document. If the highlighted sections feel stretched or insufficient to a human reader, you must rewrite those portions for clarity. Your final check should focus on the deliverable itself to ensure that the evidence cited by the AI actually provides the necessary information to the end user.