AI decision audit trail
Priya Sharma — Meridian Textiles Pte Ltd
Mapped to Taiwan FSC's "Guidelines for the Use of AI in the Financial Industry" (June 20, 2024), four-stage AI lifecycle: planning & design; data collection; model building & validation; deployment & monitoring. Generated 2026-10-01 22:05 UTC from case data current as of that time.
Not a compliance determination
FSC guidance requires firms using third-party generative AI to designate personnel who "objectively and professionally manage" the risk and to keep records of AI model design and decision-making to assess fairness. This document is that record for this case — it is evidence to support a human reviewer's decision, not a substitute for one.
Stage 1
Overview
This case draws on AI assistance in two places: identifying the beneficial owners of the subject entity and how they're connected, and assessing the evidence behind the client's source-of-wealth claims. In both cases, the AI does not make the final call — it produces a confidence assessment and a written explanation, and a human reviewer decides what happens next. Every AI response is captured in a fixed, consistent format specifically so it can be logged and reviewed here, rather than left as free-form text. AI model used: Claude, developed by Anthropic.
Stage 2
Data collection
Every document submitted for this case is kept exactly as received and is the only basis for any AI-derived finding that follows from it. 6 document(s) submitted to date.
ACRA BIZFILE — BUSINESS PROFILE (Extract), pulled 2024-11-04 by the bank using UEN 201512345K (provided by the client at onboarding) ------------------------------------------------------------ UEN: 201512345K Entity Name: MERIDIAN TEXTILES PTE. LTD. Entity Type: Local Company Business Constitution: Company Limited by Shares Company Type: Private Company Limited by Shares Date of Registration/Incorporation: 2015-03-12 Entity Status: Live Registered Address: 16 Kallang Place, Textile House, Singapore 339157 Primary SSIC: 13990 — Other Textile Manufacturing N.E.C. No. of Officers: 3 ------------------------------------------------------------ Shareholders of record (per Bizfile's Company Shareholders Details): Priya Sharma — 150,000 shares (15.0%) Sharma Holdings Pte Ltd — 650,000 shares (65.0%) Daniel Koh — 80,000 shares (8.0%) Other shareholders — 120,000 shares (12.0%) Total issued shares: 1,000,000 ordinary shares
identity verification confirms National ID S8512345A, date of birth 14 February 1985, for client Priya Sharma — matches the record on file.
Identifiers were withheld before this was sent to the AI
identity verification confirms National ID [NATIONAL-ID-REDACTED], date of birth [DATE-OF-BIRTH-REDACTED], for client Priya Sharma — matches the record on file.
date of birth: matched Priya Sharma on file
national id: matched Priya Sharma on file
ACRA BIZFILE — BUSINESS PROFILE (Extract), pulled 2024-11-04 by the bank using UEN 201811987H (Sharma Holdings Pte Ltd, identified from the Meridian Textiles shareholder register above) ------------------------------------------------------------ UEN: 201811987H Entity Name: SHARMA HOLDINGS PTE. LTD. Entity Type: Local Company Entity Status: Live ------------------------------------------------------------ Shareholders of record (per Bizfile's Company Shareholders Details): Priya Sharma — 100,000 shares (100.0%) Total issued shares: 100,000 ordinary shares
an individual named Daniel Koh is registered as holding an 8% stake in Meridian Textiles Pte Ltd, but is reported by industry contacts to be acting as a nominee for Priya Sharma — this has not been confirmed through primary documentation.
ACRA deregistration filing, 2018: confirms Priya Sharma sold her stake in a prior textile trading business for SGD 12,000,000, consistent with the claimed sale. Employer HR records confirm salary payments to Priya Sharma consistent with the claimed role and period, 2014-2019. Brokerage statements show transfers into Priya Sharma's account in 2019-2021, but the dates precede the confirmed 2018 stake sa
Employer-submitted income records, 2015-2019: salary payments consistent with the client's claimed salary income and savings accumulation over the period.
Stage 3
Model building & validation
This system does not train or customize its own AI model — it uses a general-purpose model provided by a third party (Anthropic). Taiwan's FSC guidance is explicit that firms in this position must still actively manage that risk rather than accept the AI's output uncritically. To do that, every AI output passes through the checks below before anything is added to the case record:
A new name is only linked to an existing person or company when the AI positively confirms it's the same one.
Otherwise it's treated as new and kept separate. This is what lets 'Lin Wei-Chen' — a different ordering of the same name — correctly match the existing 'Wei-Chen L.' record instead of being logged as a different person.
A known ownership percentage can never be silently erased by a later, less specific document.
If a new document doesn't state a specific percentage for a relationship that's already on file, the existing figure is kept rather than cleared. This protects a confirmed stake from being lost just because a later document mentioned the relationship without repeating the number.
Indirect ownership through a chain of companies is calculated independently, not estimated by the AI.
The AI is only asked to report direct relationships a document actually states — for example, 'Company A owns 70% of Company B.' Any indirect stake reachable through a chain of companies is then calculated separately by the system itself, so the same ownership can't be counted twice.
Every confidence score comes with a written reason, and the system leans cautious.
The AI is never allowed to return a bare number — each score must include a short explanation. It's also instructed to weigh the two directions differently: evidence that contradicts a claim should lower confidence sharply, while evidence that supports it should raise confidence more gradually. Missing a real problem is treated as more costly than being overly cautious.
National ID numbers and dates of birth are never sent to the AI model.
When a submitted document contains one, it's compared locally, within our own systems, against the record already on file — the AI is only ever told whether that comparison matched, not the number itself. This keeps the raw identifier from ever reaching the third-party AI provider, while still letting the system use a positive match as supporting evidence when resolving a name variant.
Stage 4
Deployment & monitoring
Beneficial ownership — current state
Company Act — 10% threshold
80.0% confirmed
AML Act — 25% threshold
88.0% potential
Escalated for human review
Potential aggregate control (88.0%) crosses the 25% UBO threshold pending confirmation of inferred relationships. No automated action closes this case.
| Relationship | Status | Rationale | Recorded |
|---|---|---|---|
| Priya Sharma → Meridian Textiles Pte Ltd (15% direct shareholder) | verified | ACRA Bizfile shareholder record lists Priya Sharma with 150,000 shares (15.0%) in Meridian Textiles Pte Ltd. | 2026-08-14 05:58 |
| Sharma Holdings Pte Ltd → Meridian Textiles Pte Ltd (65% direct shareholder) | verified | ACRA Bizfile shareholder record lists Sharma Holdings Pte Ltd with 650,000 shares (65.0%) in Meridian Textiles Pte Ltd. | 2026-08-14 05:58 |
| Daniel Koh → Meridian Textiles Pte Ltd (8% direct shareholder) | verified | ACRA Bizfile shareholder record lists Daniel Koh with 80,000 shares (8.0%) in Meridian Textiles Pte Ltd. | 2026-08-14 05:58 |
| Priya Sharma → Sharma Holdings Pte Ltd (100% direct shareholder) | verified | ACRA Bizfile extract shows Priya Sharma holds 100,000 of 100,000 issued shares (100%) in Sharma Holdings Pte Ltd. | 2026-08-14 05:58 |
| Priya Sharma → Daniel Koh (beneficial interest via nominee (unconfirmed)) | inferred | Industry contacts report Daniel Koh's 8% registered stake in Meridian Textiles Pte Ltd is held as a nominee for Priya Sharma; not confirmed by primary documentation. | 2026-08-14 05:59 |
How the ownership structure was uncovered
Source-of-wealth claims — how confidence changed over time
Sale of prior business stake, 2018
S$12,000,000 claimed
| When | Confidence | Status | Note |
|---|---|---|---|
| 2026-08-14 05:57 | 0% | pending | No evidence submitted yet. |
| 2026-08-14 05:59 | 80% | corroborated | ACRA deregistration filing is an official third-party government record confirming sale of the prior textile trading stake for SGD 12,000,000 in 2018, directly corroborating the claim. |
Salary income, textile trading, 2014–2019
S$9,000,000 claimed
| When | Confidence | Status | Note |
|---|---|---|---|
| 2026-08-14 05:57 | 0% | pending | No evidence submitted yet. |
| 2026-08-14 05:59 | 55% | corroborated | Employer HR records support salary payments consistent with claimed role and period, but as employer-provided (not fully independent third-party like a tax authority) records, confidence is raised only moderately. |
| 2026-08-15 16:58 | 68% | corroborated | Employer-submitted income records for 2015-2019 corroborate the claimed salary income and savings accumulation, providing third-party support beyond client self-reporting. Confidence raised but capped moderately since employer payroll records, while independent of the client, are not a fully authoritative government/tax filing. |
Brokerage account transfers, 2019–2021
S$6,000,000 claimed
| When | Confidence | Status | Note |
|---|---|---|---|
| 2026-08-14 05:57 | 0% | pending | No evidence submitted yet. |
| 2026-08-14 05:59 | 20% | downgraded | Brokerage statements show transfers in 2019-2021, but the document notes some transfer dates precede the confirmed 2018 stake sale, creating a timeline inconsistency that undermines the claimed source of these funds; needs clarification before further corroboration. |
Aggregate check — claimed sources vs actual assets
Claimed sources total S$27,000,000 against total assets of S$32,000,000. A gap of S$5,000,000 has no claimed source, independent of any single claim's status.