deepidv
All AI Prompts
FinTechReview Prompt

Compliance Prompt to Audit AUSTRAC Suspicious Matter Report Quality

This **Luna** review prompt takes your draft **suspicious matter report (SMR)** queue and AUSTRAC's **2026-27 supervisory priorities**, then scores each report before it is filed. Luna, the deepidv compliance overseer, checks **filing-window timeliness** against the three-business-day rule (24 hours for terrorism financing), rates **grounds-for-suspicion narrative quality** against AUSTRAC guidance, confirms the attached **customer record** is current with ongoing due diligence status, checks internal consistency across the narrative, transaction log, and evidence bundle, and flags **typology mismatches** where described conduct matches a documented pattern. Built for Tranche 2 reporting entities, accountants, conveyancers, trust and company service providers, and virtual asset service providers, who must survive an AUSTRAC supervision sweep with SMRs that are timely, complete, and coherent.

Compliance Prompt to Audit AUSTRAC Suspicious Matter Report Quality

How to use this prompt

  1. 1

    Open Luna in the deepidv dashboard and paste the full prompt, or run it in Claude, ChatGPT, or Gemini if you are auditing SMR quality outside the platform.

  2. 2

    Replace the INPUT section with your draft SMR queue, the attached customer records and transaction logs, and the filing timestamps for each report.

  3. 3

    Run the prompt and read the ranked queue first: it surfaces any SMR past 50 percent of its filing window that is still in draft.

  4. 4

    Hand the per-report fix list to the drafter and escalate anything within 12 hours of its deadline to your compliance officer.

  5. 5

    Re-run the prompt on every SMR before filing so the program AUSTRAC examines is the one you actually run.

The prompt

Luna, act as an AUSTRAC SMR quality auditor for a Tranche 2 reporting entity.

Context: AUSTRAC's 2026-27 supervisory priorities target suspicious matter reports that are late, incomplete, or lack a coherent grounds-for-suspicion narrative. Filing windows are 3 business days, or 24 hours where terrorism financing is suspected.

For each draft SMR in my review queue:
1. Score timeliness: hours remaining in the filing window, and flag anything past 50% of the window still in draft.
2. Score narrative quality against AUSTRAC guidance: does the report state who, what, when, the specific grounds for suspicion, and the supporting customer identification and transaction evidence?
3. Verify the customer record attached is current: last verification date, ongoing due diligence status, and any unresolved re-verification triggers.
4. Check internal consistency: amounts, dates, and identifiers must match across the narrative, the transaction log, and the evidence bundle.
5. Flag typology mismatches: if the described conduct matches a documented typology (structuring, mule activity, synthetic identity), confirm the typology is named and the matching indicators are listed.

Output: a ranked queue with a quality score per SMR, the specific fixes required before filing, and a daily summary I can table at the compliance committee. Escalate anything within 12 hours of its deadline. Where a report cannot be scored from the input, flag it as an open question instead of guessing.

Test it in Claude or another LLM

This prompt is built for the Luna agent inside deepidv, where Luna reads a firm's live SMR queue and customer records against AUSTRAC's supervisory priorities. You can dry-run the same workflow in any general LLM first with synthetic SMR drafts to see the scoring shape before pointing it at real reports.

  1. 1

    Paste the full prompt into Claude, ChatGPT, or Gemini, but replace the opening 'Luna,' with a role instruction such as 'Act as an AUSTRAC SMR quality auditor for a Tranche 2 reporting entity.' Keep the OUTPUT sections exactly as written.

  2. 2

    Under the INPUT section, paste the synthetic sample block below so the model has draft SMRs, filing timestamps, and customer records to evaluate.

  3. 3

    Add one framing line: 'This is synthetic test data. Where a report cannot be scored from the input, flag it as an open question instead of guessing.'

  4. 4

    Check the output shape: a ranked queue with a quality score per SMR, the specific fixes required before filing, and a daily committee summary. If any score invents evidence the input does not contain, tighten the role line and re-run.

  5. 5

    Once the output shape is right, run it live in the deepidv dashboard where Luna scores your real SMR pipeline.

Synthetic sample data to paste alongside the prompt

Fake test data, safe to share with any LLM. Swap in your own once the output looks right.

DRAFT SMR QUEUE (synthetic, fake):
- 4 draft SMRs, refs SMR-TEST-01 through 04, mix of structuring and third-party-deposit typologies
- Filing clocks (fake): SMR-01 opened 62h ago (3-day window), SMR-02 opened 20h ago (terrorism-flag, 24h window), SMR-03 and 04 opened 8h ago
CUSTOMER RECORDS (fake): SMR-01 customer last verified 14 months ago, ODD overdue; others current
TRANSACTION LOGS (fake): amounts and dates provided per SMR; SMR-03 narrative amount does not match log
AUSTRAC PRIORITY TAGS (fake): timeliness, narrative quality, typology naming
OPEN ITEM (fake): whether SMR-02 grounds meet the 24h terrorism-financing threshold

FAQ

What is AUSTRAC scoring in an SMR quality review?

AUSTRAC's 2026-27 priorities target suspicious matter reports that are late, incomplete, or lack a coherent grounds-for-suspicion narrative. A quality review checks whether each report states who, what, and when, names the specific grounds for suspicion, attaches current customer identification and transaction evidence, and is filed inside the three-business-day window, or 24 hours where terrorism financing is suspected.

How does this prompt help before an AUSTRAC supervision sweep?

It scores every draft SMR before filing, so timeliness slips, thin narratives, stale customer records, and typology mismatches are caught while they are still fixable. AUSTRAC has flagged SMR quality and timeliness as a supervisory priority, so a documented pre-filing audit trail is exactly the evidence an examiner asks for.

Can I use this prompt outside the deepidv dashboard?

Yes. The structure works in Claude, ChatGPT, or Gemini as an SMR quality-audit framework and returns the ranked queue, per-report fixes, and a committee summary. Live scoring against your real SMR pipeline and customer records only runs when it executes inside the deepidv dashboard through Luna.

Run it with live verification data

These prompts work in any LLM. Inside the deepidv dashboard, Luna, Arbiter, and Arc run them against your real sessions, screening lists, and audit trails.

Book a Demo