小論文作成アドバイザー-AI essay drafting tool

Overview of 小論文作成アドバイザー (Concise Essay / Promotion Paper Advisor)

小論文作成アドバイザー is a specialized writing-advisor workflow and assistant designed to produce, refine, and validate Japanese-style argumentative essays and promotion papers (昇格論文) with rigorous structure, measurable goals, and audit-ready citations. Its design purpose is fourfold: (1) enforce reproducible output structure required by Japanese corporate and academic contexts (summary → outline → sectioned body → conclusion → appendix), (2) operationalize argument quality by linking claims to KPIs and customer-value hypotheses, (3) provide staged proofreading and measurable grading so authors can iterate efficiently, and (4) generate deliverables that match submission requirements (voice:である調 when requested, paragraph metadata, APA7 citations, tables/figures and RACI roadmaps). Core behavioral rules and guarantees (examples): • Draft sizing rule: when a user specifies a target length (e.g., 1,000 words), the system produces a draft at 1.20–1.30× that length (example: 1,200–1,300 words) to leave space for compression and enrichment; in the final step the assistant trims the text to within ±5% of the original target and documents what was removed and why. Example: requested 1,000 words → initial draft 1,260 words → final trimmed output 995 wordsIntroduction to 小論文作成アドバイザー with a short deletion log. • Paragraph metadata: every paragraph ends with a metadata tag summarizing local counts and cumulative totals, shown as "〔paragraph_chars / section_cum_chars / body_cum_chars〕" so reviewers and editors can quickly verify length and structure. Example: paragraph ending tag: "〔482 / 1,932 / 4,280〕". • KPI-first framing: the assistant selects up to three primary KPIs from four categories (Outcome/成果, Efficiency/効率, Economics/経済性, People/人), and for each KPI provides baseline → target → concrete deadline and measurement method. Example: "Customer NPS baseline 42 → target 55 by 2026-03-31; measurement: quarterly NPS survey". • Customer-value hypothesis: for each primary and secondary customer the assistant creates a Before/After quantified hypothesis (e.g., before: average task completion 5.2 days → after: 3.1 days; expected 40% time saved), with the data source or how to measure it. • Citation & traceability: external sources are formatted in APA7 plus a footnote containing the URL and retrieval date; internal references are recorded as "document name / department / version / date". Example APA7 template provided: "Author, A. A. (Year). Title. Publisher. URL (Retrieved YYYY-MM-DD)". Practical scenario that demonstrates core design: A mid-level manager must submit a 1,200-word promotion paper arguing for running a pilot product line. The assistant: 1) produces a 1,440–1,560-word draft (1.2–1.3×), including summary (200–300 words) with KPIs up-front, an outline, body paragraphs each ending with metadata tags, Table 1 (cost/benefit sensitivity), Figure 1 (quarterly roadmap with RACI), and APA7 citations; 2) runs L1→L2→L3 proofreading and returns a revision table showing changes and rationale; 3) runs grader-mode (100-point rubric) and flags paragraph vulnerabilities (e.g., insufficient evidence in paragraph 4) with replacement text; 4) trims the draft to 1,200±5% (1,140–1,260 words) and attaches a deletion log explaining what was removed and why. This produces a submission-ready document that meets corporate formatting and evidentiary expectations.