Units
1. Introduction to Artificial Intelligence in Finance and Tax: Roles, Boundaries, Verification and Ethics 2. Tax Legislation Research: Law, Communiqué, Ruling and Judicial Decision 3. Declaration Preparation Support: VAT, Withholding Tax, Provisional and Corporate Tax 4. Tax Calculation and Base: Recalculation Discipline 5. Financial Statement Analysis: Balance Sheet, Income Statement and Ratio Interpretation 6. Tax Compliance Checklists and Process Automation 7. Notification, Objection and Correspondence Drafts: Tax Office and Reconciliation 8. Tax Review and Audit Preparation: Document, Evidence, Defense 9. Transfer Pricing, VAT Refund and Complex Issues 10. Customer Data and Privacy: KVKK, VUK Confidentiality and Secure Tools 11. Managing the Risk of Errors, Hallucinations and Misinterpretations 12. End-to-End Secure Workflow: Quality Control, Expert Approval and Automation
Unit 8 / 12

Tax Review and Audit Preparation: Document, Evidence, Defense

Gains:

  • Document compilation, amount reconciliation and pre-mapping of possible criticism points with artificial intelligence in preparation for tax audit
  • Ability to use artificial intelligence in the production of defense arguments and scenario drafts and support each argument with legislation and documentary evidence
  • Being able to distinguish that the final defense and declaration during the examination process is the responsibility of the professional and the taxpayer, and that artificial intelligence is only a preparation tool.

Tax audit is the administration's inspection of a taxpayer's books, documents and statements for compliance with the legislation. For the professional, this is the process with the highest tension but with the most preparation. A well-prepared file prevents an unfair assessment; An unprepared file leaves even a decent application vulnerable. AI is a powerful preparation tool here: it organizes document compilation, reconciles amounts, pre-maps possible points of criticism, drafts defense arguments. But the principle is immutable: final defense and declaration are the responsibility of the professional and the taxpayer; AI will only speed up preparation.

The logic of audit preparation

Good preparation answers three questions in advance: What will the administration look for? What is our position? With what document do we prove this? Artificial intelligence helps unpack these three questions systematically.

The main components of the preparation:

  • Document compilation: Invoices, contracts, bank and book records of the relevant period are collected completely.
  • Reconciliation: It is ensured that the declared amounts are consistent with the records and documents.
  • Risk/criticism map: It is predicted in advance which item the administration may criticize.
  • Defense argument: A position based on legislation and documents is prepared against every possible criticism.
  • Scenario rehearsal: Possible questions and answers are practiced in advance.
Tip: Ask the artificial intelligence, "If you were a tax inspector, which items would you query in this file?" ask. The model is very useful in mapping points of criticism; but evaluate each point with your own professional judgment and documentation.

Safe preparation step by step

  1. Determine the scope. Which period and which tax type does the review cover?
  2. Organize documents. Have the AI ​​produce a document checklist; actually find each document.
  3. Make the agreement. Question the consistency of the declaration-record-document trio; Recalculate the differences.
  4. Map points of criticism. List and prioritize possible objection issues with the model.
  5. Set up defense. Draft legislation + documentary evidence for each criticism; Confirm the basis.
  6. Try to rehearse. Prepare question and answer scenarios; The final position is determined by the professional.

The role and limit of artificial intelligence

AI is an accelerator in preparation; is not a party to the investigation itself. The model does not respond to the administration, conduct the defense, or produce the document on your behalf. Even if he produces a defense argument, every item on which that argument is based must be confirmed and every document must actually exist. "Artificial intelligence argued like this" is not a valid justification in the review.

preparation mission

Artificial intelligence contribution

human responsibility

Document list

Comprehensive checklist

Actual existence of the document

consensus

Difference detection, question generation

Reason for difference and correction

Criticism map

List possible objections

Assessing real risk

defensive argument

Draft justification

Confirmation of basis, final position

Response to administration

None

Fully professional

three mini cases

Case 1 — Previously captured difference. A consultant reconciles records with a statement to the AI ​​prior to review; The model indicates that there is an unexplained difference of 128,000 TL in an account. The consultant investigates the difference, finds that there is a recording error, and corrects it. When the review comes, the file is clean. Lesson: artificial intelligence makes the difference visible, humans find the solution.

Case 2 — Fabricated defense basis. An intern prepares the defense produced by artificial intelligence against a criticism; There is actually no reference to notification in the defense. On review, this weakens the defense. The correct way was to confirm every basis and eliminate the fabricated attribution.

Case 3 — Good handling. In a large file, a CPA tells the AI ​​to “extract the 10 most likely points of criticism through an inspector's eye and tell me what document I need for each.” The model produces a comprehensive map. CPA evaluates each point with legislation and documents, prioritizes real risks and builds the defense file accordingly. Preparation takes days instead of weeks, and no critical point is missed.

Weak prompt / Strong prompt

Weak prompt:

Defend me against tax audit, will you justify whatever I say?

The model produces a text that is biased, unsupported, and full of fabricated justifications.

Powerful prompt:

Your role: tax audit preparation assistant.Task: List the items an INSPECTOR might question based on the file summary below and write down what documentation/basis for defense is required for each.Rules:- Don't take sides; Show both the risk and the possible defense objectively. - If you add grounds, mark "must be confirmed"; Fabricating article/notification.- Specify the document required for each defense. Format: point of criticism | risk level | defensive direction | required document.[FILE SUMMARY (anonymised): ...]

The second claim makes the model an objective risk mapper; binds the defense to documentation and confirmation.

For the reconciliation query:

Compare the declaration amounts with the record amounts below and question whether there are any unexplained differences. List the differences, mark possible causes as "needs investigation"; giving definitive results.[DECLARATION AND REGISTRATION SUMMARY: ...]

For document checklist:

Generate the document checklist required for this review scope: what documents should be ready by period, tax type and item? Mark general items as "confirm by company". Scope: [period, tax type]

For defense rehearsal:

For the following point of criticism, write down 5 difficult questions an inspector might ask and a sample response framework to verify the basis for each. Establishing a definitive legal claim; Give the framework, I will confirm the basis.[CRITICISM POINT: ...]

Common mistakes

  • Having the model do the defense. The final position and basis lies with the professional; produces model drafts.
  • Using a fitting support. Not every reference in the defense is included in the file without being confirmed.
  • Assuming the document. Even though the model says a document "should" exist, it must actually exist.
  • Ignoring the difference. Any unexplained difference in consensus should be investigated.
  • Losing objectivity. A biased “prove me right” approach hides real risk.

In summary

Tax examination comes with preparation, and AI powerfully accelerates this preparation: it is valuable in document compilation, memorandum, critique point mapping, and defense drafting production. However, the final defense and declaration are the responsibility of the professional and the taxpayer. Every defense argument must be supported by legislation and documentary evidence, every basis must be confirmed, every document must actually exist. AI is a preparedness tool; is not a party to the investigation.

Application task

Create a sample review file (period, tax type, a few risky items). Get a critique point map and document list from AI with the powerful prompt template. Confirm the required basis for each point of criticism from the official source and check the actual availability of the required document. Check and note whether there is a fabricated basis in the defenses produced by the model.

checklist

  • [ ] I clarified the scope of review (period, tax type).
  • [ ] I pulled out a document checklist and actually found every document.
  • [ ] I reconciled the declaration-record-document and recalculated the differences.
  • [ ] I mapped possible points of criticism and prioritized them according to actual risk.
  • [ ] I have confirmed each defense basis from the official source.
  • [ ] I determined the final defense position using my own professional judgment.
  • [ ] I protected privacy by anonymizing file data.
Previous Next