Business Finance - Accounting Conference Submission Assignment (Due in 15h)

AISC_Researcher_Prompt_Kit.pdf

Home >Business & Finance homework help >Accounting homework help >Business Finance - Accounting Conference Submission Assignment (Due in 15h)

AISC A* Manuscript Co-Researcher

Researcher Prompt Kit

Generation, Audit, Audit Response, and Submission Prompts

Purpose of the Kit

This kit contains the prompts a researcher pastes into the AISC A* Manuscript Co-

Researcher to move a study from raw materials to a submission-ready manuscript. The GPT

already holds the AISC template, the Operating Instructions, and the Professional Academic

Voice Master Prompt as knowledge files; the prompts below tell it which of those procedures to

run, with which inputs, and in what order. Each prompt is self-contained, so a researcher can

return to the cycle at any point in a new chat by attaching the latest manuscript and the relevant

records.

The cycle has four steps (Table 1). Prompt 1 generates or refines the manuscript section

by Section and builds the Word file. Prompt 2 audits the manuscript and produces a structured

audit report without rewriting anything. Prompt 3 takes the audit report back into the GPT,

validates and triages every finding, revises the manuscript where the Evidence permits, and

returns a revised manuscript with an Audit Response Matrix. Prompt 4 performs the final

submission checks. Prompts 2 and 3 repeat until the audit certifies the manuscript as ready,

normally within two cycles.

Table 1: The Four-Step Manuscript Cycle

Ste

Prompt Researcher attaches GPT returns Researcher

decides

1 Manuscript

Generation

Sources (full text),

archived output, Excel

output, draft (Refine

Mode), contribution

statement

Sections for

confirmation; Content

Preservation Ledger;

Citation Summary

Table; Word

manuscript

Confirms each

Section; supplies

missing Evidence

2 Manuscript

Audit

Word manuscript;

archived output;

sources where available

Audit report with

severity, anchor,

evidence quotation, and

owner for every finding

Reviews the

report; prepares

decisions on

researcher-owned

findings

3 Audit

Response and

Revision

Word manuscript; audit

report; Evidence and

decisions for

researcher-owned

findings

Revised Word

manuscript; Audit

Response Matrix;

Revision Log; re-audit

certification

Accepts revisions

or returns to Step

4 Submission

Readiness

Final Word manuscript;

venue requirements

Submission checklist;

optional separate title

page and AI-use

disclosure draft

Verifies citations

personally;

submits

Three rules hold throughout the cycle. The GPT never invents a source, statistic,

quotation, or DOI to close a gap or resolve an audit finding; it inserts a marker and asks. The

GPT never shortens the manuscript by deleting substantive content; it condenses wording,

relocates secondary material, and proves preservation through the Content Preservation Ledger.

The researcher remains responsible for every citation, every number, and the contribution

statement, and personally verifies every citation against its original source before submission.

Before You Start

Prepare the following before pasting Prompt 1. Gaps are acceptable because the GPT marks what

is missing, but every gap becomes a marker that must be resolved before submission.

1. The full texts (PDFs) of the sources the manuscript will cite, or the existing reference list

with the PDFs of the core sources.

2. The complete archived statistical output (Stata log, R output, or SPSS output), not a

summary of it.

3. The Excel workbook containing the tables you will paste back into the manuscript, with

each table on a named sheet.

4. The existing draft, if you are using Refine Mode, as a .docx file.

5. A contribution statement of four to six sentences that you have written yourself; the GPT

inserts it verbatim.

6. The target journal's author guidelines, if you are writing for a named journal in addition to

the AISC.

Fill in every bracketed field in a prompt before pasting it, and delete lines that do not

apply. Leave the capitalized labels in place; the GPT uses them to locate each instruction.

Prompt 1. Manuscript Generation

Use Prompt 1 at the start of a new manuscript (Develop Mode) or when bringing an

existing draft into the AISC template (Refine Mode). By default, the GPT delivers one Section at

a time and waits for your confirmation; choose the full-draft option only when the Evidence is

complete, and you prefer to review the whole manuscript at once.

PROMPT 1: MANUSCRIPT GENERATION

Apply the Professional Academic Voice in full, and follow the AISC template and your

Operating Instructions for every step.

STUDY PROFILE

Working title: [title]

Mode: [Develop Mode: draft from the attached materials / Refine Mode: refine the attached

draft]

Destination: [AISC only / AISC and [journal name], author guidelines attached]

Discipline and subfield: [e.g., accounting, corporate governance]

Problem and its consequence: [two or three sentences]

Theoretical lens: [theory or theories]

Core mechanism: [one or two sentences stating what the theory predicts and why]

Research questions: RQ1 [ ]; RQ2 [ ]

Hypotheses (if already specified): H1 [ ]; H2 [ ]

Data: [source, period, unit of analysis, sample filters; anonymize the provider and site if

identifying]

Design and estimator: [e.g., panel regression with firm and year fixed effects; Stata 18]

Significance threshold for characterizing hypotheses: [e.g., p < .05, two-tailed]

Contribution statement (insert verbatim, do not edit): [four to six sentences written by the

researcher]

MATERIALS ATTACHED

[List each file: draft manuscript, source PDFs, archived output, Excel workbook, conceptual

model figure]

TABLES

Tables from Excel (insert slots only, never retype): Table [n], [title], [workbook, sheet]; Table

[n], [title], [workbook, sheet]

Tables to build in Word from the archived output: [list, or "none"]

LENGTH

Target: 25 pages from the Abstract to the Conclusion, using the Section ranges in your page

budget, with Results prose of 3.5 to 4.5 pages and the remaining Results allocation reserved

for the Excel tables.

If the material or draft exceeds the budget, apply the Lossless Compression Protocol: build the

Content Preservation Ledger before any reduction, use permitted techniques only, relocate

secondary material to the Appendixes with pointers, and reconcile the ledger after each

Section. Delete no finding, statistic, citation, hypothesis, table, figure, robustness test, or

limitation. If the budget cannot be met without loss, stop and report the excess by Section. You

may also ask the researcher for permission to exceed the results pages.

TASKS

1. Confirm the setup in one message: voice, Mode, destination, materials received, materials

missing, and the Excel table slots you will create. In Refine Mode, number every paragraph of

the draft by Section and report the counts, and deliver the Content Preservation Ledger.

2. Draft or refine the manuscript one Section at a time in template order. In Develop Mode,

draft the Abstract last.

3. For each Section, deliver: integrity status; the section text under its template heading; a short

revision note; the length line (words, estimated pages, pages reserved for Excel tables, and, if

compressed, the original word count and ledger reconciliation); the Evidence and Gap Log;

and a request for my confirmation.

4. Cite only the attached or confirmed sources. Copy every number verbatim from the archived

or Excel output. Use markers wherever Evidence is missing, and label your own suggestions

[AI-PROPOSED: CONFIRM].

5. Apply the double-blind review rules throughout: mask institutions, sites, approval numbers,

and providers, and cite my own prior work in the third person.

6. After the Conclusion, deliver the References in APA 7, the Citation Summary Table, the

Appendixes, and the final ledger reconciliation.

7. Run the Pre-Delivery Audit and report it, then build the Word manuscript with the template

build script and report the verification output.

CONFIRMATION MODE

[Section by Section (default) / Full draft in one pass, then a single confirmation]

Continuation Prompts for Step 1

Short prompts keep the section workflow moving without repeating Prompt 1.

Table 2: Continuation Prompts Used During Generation

Situation Prompt to paste

Section accepted Confirmed. Lock [Section] as version [n] and proceed to [next

Section].

Section needs

changes

Revise [Section] only, as follows: [instructions]. Keep every ledger

item, citation, and number. Report which paragraphs changed and

confirm that the six connections still hold.

Evidence now

supplied

The attached [source/output/value] resolves markers [list anchors].

Replace those markers only with the supplied Evidence, and report

each replacement.

Suggestion rejected Reject [AI-PROPOSED item] because [reason]. Remove it and record

the decision in the Evidence and Gap Log.

Section too long Compress [Section] to [n] pages under the Lossless Compression

Protocol. Show the ledger reconciliation and the original and new

word counts.

Ready to build All sections are confirmed. Run the Pre-Delivery Audit, then build the

Word manuscript from the template and report the verification output.

Prompt 2. Manuscript Audit

Use Prompt 2 on the Word manuscript produced in Step 1, preferably in a new chat so the

drafting conversation doesn't influence the audit. The audit reports findings only; it does not

rewrite the manuscript. The Owner column tells Prompt 3 which findings the GPT may resolve

and which require you. If you use your separate A* Audit Framework instead, ask it to add the

same Owner column so that its report can enter Prompt 3 unchanged.

PROMPT 2: MANUSCRIPT AUDIT

Audit the attached manuscript under the Pre-Delivery Audit in your Operating Instructions

and, because the Professional Academic Voice applies, the variation diagnostics in its Part IV.

Do not rewrite, correct, or propose replacement text. Do not propose corrected statistical

values.

MATERIALS ATTACHED

Manuscript: [file name].docx

Archived statistical output: [file names, or "not supplied"]

Excel table output: [workbook name, or "not supplied"]

Source PDFs: [list, or "not supplied"]

Content Preservation Ledger from Step 1: [attached / not applicable]

TARGET

[AISC only / AISC and [journal name]]

AUDIT SCOPE

Module A: template conformance (headings, order, styles, keywords, page breaks, table slots).

Module B: six-connection check, each connection PASS or FAIL.

Module C: page budget by Section, Excel table reservation, and ledger reconciliation.

Module D: double-blind compliance, including self-citation and document properties.

Module E: integrity (open markers, numbers against the supplied output, hypothesis

characterization, causal language, findings outside their sections).

Module F: voice compliance (locked rules, contamination patterns, restricted constructions,

variation diagnostics); run the voice scan on each Section.

EVIDENCE RULES

1. Number every paragraph by Section first and report the counts.

2. Support every finding with a verbatim quotation of five to twenty words and an anchor in

the form §4 ¶2 S3. Omit any finding you cannot quote.

3. Classify every check by tier. Report Tier 2 checks as not assessed where sources were not

supplied. Never state that plagiarism was or was not detected, never give a similarity

percentage, and never confirm that a DOI resolves.

OUTPUT

1. Paragraph counts by Section.

2. Audit report table: ID | Module | Severity (Critical, High, Medium, Low) | Anchor |

Evidence quotation | Issue | Required action | Owner (AI: revisable from the manuscript and

supplied evidence / RESEARCHER EVIDENCE: needs a source, value, or output /

RESEARCHER DECISION: needs a judgment only the researcher can make).

3. Six-connection table and page budget table.

4. Certification (Ready for submission / Ready after minor revision / Not ready) with the count

of findings at each severity level.

5. A list of the Evidence and decisions I must supply before Prompt 3.

Prompt 3. Audit Response and Revision

Use Prompt 3 after reviewing the audit report. Attach the manuscript and the report, and

provide the Evidence and decisions the report requested. The GPT first verifies each finding,

then sorts the findings by owner, revises only what the Evidence permits, and returns a revised

manuscript with a matrix that accounts for every finding. A finding that needs Evidence you

have not supplied stays open; the GPT does not resolve it by rewording the claim around the gap

or by inventing support.

PROMPT 3: AUDIT RESPONSE AND REVISION

Apply the Professional Academic Voice in full. Respond to the attached audit report and revise

the attached manuscript into a submission-ready version in the AISC template.

MATERIALS ATTACHED

Manuscript: [file name].docx

Audit report: [file name, or pasted below]

Content Preservation Ledger: [attached / not applicable]

Evidence for RESEARCHER EVIDENCE findings: [list each finding ID with the source,

value, or output supplied]

Decisions for RESEARCHER DECISION findings: [list each finding ID with Accept, Revise

as follows, or Reject, and the reason]

Findings I dispute: [finding ID and reason, or "none"]

TASKS

1. Validate the report. Confirm that each evidence quotation appears at its anchor in the

manuscript. Mark any finding whose quotation cannot be located as NOT REPRODUCIBLE

and take no action on it.

2. Triage every remaining finding into one category:

A. Revisable now: voice, contamination, flow, structure, template conformance, masking,

length, and connection repairs that need no new evidence.

B. Evidence supplied: RESEARCHER EVIDENCE findings for which I attached the

Evidence above.

C. Decision supplied: RESEARCHER DECISION findings for which I recorded a decision

above.

D. Open: findings that still need my Evidence or decision.

E. Disputed: findings that conflict with the template, the Evidence, or required technical

language; give the reason with a quotation.

3. Revise categories A, B, and C in order of severity, Critical first. Change only the passages

each finding concerns and the passages a repair must reach to keep the six connections intact.

4. Preserve content. Keep every ledger item, citation, statistic, hypothesis, table, table slot,

figure, and limitation. Add no source outside the confirmed set. Change a number only where

the supplied output shows that the manuscript misreports it, and quote the output line. Insert

my contribution statement verbatim.

5. Leave category D findings unrevised and keep or insert the appropriate marker.

6. Re-audit: run Modules A to F on every revised passage, and rerun the six-connection check,

the page budget, the ledger reconciliation, the Citation Summary Table, and the voice scan on

the full manuscript. Report any new findings the revision created.

7. Build the revised Word manuscript from the template with the build script, named

[title]_AISC_Blind_Manuscript_R[n].docx, and report the verification output.

OUTPUT

1. Audit Response Matrix: Finding ID | Severity | Category | Action taken | Revised location

(anchor) | Status (Resolved / Resolved with researcher input / Open: researcher action /

Disputed / Not reproducible).

2. Revision Log: for each change, the anchor, a quotation of the original wording (five to

twenty words), and the revised wording.

3. Ledger reconciliation and Citation Summary Table after revision, with counts compared

with the previous version.

4. Re-audit certification (Ready for submission / Ready after minor revision / Not ready), with

counts by severity and the remaining researcher actions.

5. The revised Word manuscript.

Do not report a finding as resolved unless the revised text resolves it.

Repeating the Cycle

When the re-audit certification is not ready, or when you add new Evidence, run Prompt

2 again on the revised manuscript (in a new chat where possible), then Prompt 3 with the new

report. Number each revision (R1, R2) so that the matrices and logs remain traceable. If the same

finding returns after two cycles, treat it as a researcher decision rather than a drafting problem,

because the GPT cannot resolve a disagreement about content by rewording.

Prompt 4. Submission Readiness

Use Prompt 4 once the certification reads Ready for submission or Ready after minor

revision and you have accepted the remaining Low findings. Do not change the blind manuscript

in this step unless a check fails.

PROMPT 4: SUBMISSION READINESS

Run the final submission checks on the attached manuscript for [AISC / AISC and [journal

name]].

MATERIALS ATTACHED

Final manuscript: [file name].docx

Venue requirements: [author guidelines, submission system notes, or "AISC defaults"]

Excel workbook for table reinsertion: [workbook name]

TASKS

1. Confirm template conformance: the nine Level 1 headings verbatim and in order, the title,

keywords, page breaks, footer, and styles.

2. Confirm double-blind review compliance in the text and in the document properties, and list

every [ANONYMIZED ...] placeholder so that I can restore the details in the camera-ready

version.

3. List every Excel table slot with its table number, title, workbook, and sheet, in order of

appearance, and confirm that each table is introduced in the text before it appears.

4. Confirm that no marker other than [ANONYMIZED ...] remains.

5. Report the length from the Abstract to the Conclusion: prose pages, reserved table pages,

and total against the 25-page target.

6. Produce a submission checklist that I complete personally, including citation verification

against original sources, Word Inspect Document, table reinsertion, and figure resolution.

7. If I request them, prepare as separate Word files (never inside the blind manuscript): [a title

page with author names, affiliations, corresponding author, and acknowledgments, using the

details I supply / a draft AI-use disclosure statement consistent with the venue's policy / a

cover letter].

OUTPUT

The checks as a table (Check | Result | Action needed), the placeholder list, the table slot list,

the submission checklist, and any separate files requested.

After the Final Prompt

Four tasks remain with the researcher and cannot be delegated. Each citation is verified

against its source, which takes roughly three to five hours for a manuscript with 40 to 55

citations. Paste the Excel tables into their slots, then delete the slot placeholder lines and check

the page length against the budget. Word's Inspect Document function (File, Info, Check for

Issues) is run to remove comments, tracked changes, and personal information added after the

final build. Prepare AI assistance disclosure according to the conference or journal policy and

submit it where the venue requires it.

Table 3: Files Produced Across the Cycle

File or record Produced in Kept for

Content Preservation Ledger Step 1 (and updated Proof that compression and

File or record Produced in Kept for

in Step 3) revision removed no substantive

content

Citation Summary Table Steps 1 and 3 Reconciling in-text citations with

the reference list

[title]_AISC_Blind_Manuscript.

docx

Step 1 First audit

Audit report Step 2 Input to Prompt 3; record of the

review

Audit Response Matrix and

Revision Log

Step 3 Traceability of every change;

source for reviewer response

letters later

[title]_AISC_Blind_Manuscript_

R[n].docx

Step 3 Submission version after

certification

Submission checklist; optional

title page and disclosure draft

Step 4 Final personal checks and

submission files