Prompt & LLM Utilization

Generating acceptance test candidates from a fictional reservation specification using Copilot Chat. We introduce prompts that separate boundary values, error scenarios, and undefined expected results, while verifying traceability to specification numbers and handling diffs during changes.

Organizing fictitious incident notes into a timeline using Copilot Chat. We demonstrate prompts and verification examples that separate observations, performed operations, candidate causes, and recovery determinations, ensuring that a single terminal's success is not misconstrued as system-wide recovery.

Reviewing omissions in handover notes using Copilot Chat. This guide explains how to identify prerequisites, decision criteria, and stopping conditions from a fictional return procedure, and how to convert them into questions for the administrator without fabricating non-existent operations or contacts.

Creating FAQs with Copilot Chat based on a fictional equipment lending guideline. This article introduces prompts to separate application conditions and exceptions, source lines, and questions not covered in the guidelines, as well as a review method to prevent conditions from being dropped from concise answers.

Comparing hypothetical proposals with differing monthly, annual, and initial costs using Copilot Chat. We explain how to review items before ranking them by using prompts that standardize cost units, user counts, and tax conditions while retaining calculation formulas and missing conditions.

How to use Copilot Chat without confusing execution, approval, references, and agreements in weekly reports. It demonstrates organizing a fictional draft memo into four categories, a prompt that retains rationale and uncompleted items, and a comparison procedure to change states.

Organizing the target, deadline, and location of materials in request emails using Copilot Chat. Using a fictional email and a copy-paste prompt, we test extracting missing conditions, formulating confirmation questions, and drafting replies without adding unconfirmed commitments.

Introducing a Copilot Chat prompt that uses fictional meeting notes to separate decisions, proposals, and unassigned items. It explains verification procedures to cross-check source lines, deadlines, and completion criteria without arbitrarily increasing tasks.