Government

Protecting national collateral through resilient fiduciary registration

Protecting national collateral through resilient fiduciary registration

Jenis Proyek

Ministry of Law and Human Rights

team

Lacak.io

year

2025

role

Lead Product Designer · Lacak.io

Jenis Proyek

Ministry of Law and Human Rights

year

2025

team

Lacak.io

role

Lead Product Designer · Lacak.io

Monday Merch Platform 2.0 interface for configuring branded merchandise orders and delegating payment to finance.


Designed under Lacak.io for the Ministry of Law and Human Rights (Kemenkumham RI / Ditjen AHU). Serving as the transaction ledger securing over $26B+ (IDR 400T+) in national banking & multifinance credit and generating $70M+ (IDR 1.1T) in annual non-tax state revenue (PNBP), this case study analyzes how we eliminated the 30-day statutory deed expiration trap and cut corporate filing drop-offs by 35%.

THE STAKES

The 30-day legal cliff: when a system error means a lawsuit

In consumer software, a crash causes user churn. In national legal infrastructure, a registration failure strips banks of their security rights over billions in collateral assets.

The statutory 30-day dead-end (Article 4, PP No. 21/2015)

Under Indonesian statutory law, whenever commercial banks or financing institutions issue loans against movable collateral (vehicles, heavy machinery, inventory), the notary has exactly 30 calendar days from signing the deed to register it in the AHU database. If that 30-day window closes without a completed registration, the filing exceeds the statutory registration window, creating a compliance and re-filing risk.

A failed registration can prevent the lender from establishing preferential creditor rights (Hak Preferen), turning secured credit into high-risk unsecured debt. The notary faces civil malpractice lawsuits. In the legacy portal, if a notary uploaded a batch of 50 vehicles and just one engine serial number had a typo, the entire submission was rejected and wiped out.

Registration deadline: PP No. 21/2015, Article 4 (BPK RI)

Syndicated lending block

When multiple banks jointly funded infrastructure equipment, the legacy single-creditor model failed completely, forcing manual workarounds that breached legal auditability.

SIMPONI voucher deadlocks

Clerical filing was tightly coupled with state non-tax billing. If treasury payment was delayed, the entire draft expired and could not be reopened.

BEHAVIORAL DYNAMICS

The triple-stakeholder pressure matrix

Public sector platforms fail when treated as generic data forms. We mapped the daily operational environments, mental models, and vulnerabilities of all three parties involved in every collateral transaction.

The notary clerk & drafter

• Environment: Busy law firm drafting 30+ deeds daily under the threat of professional civil liability.
• Mental Model: Treats registration as a legal compliance checklist. Fears that a single typo in a 17-character VIN will void the legal deed.
• Behavioral Vulnerability: When an error wipes out 40 minutes of data entry, panic sets in as the 30-day statutory clock ticks down.

Core Need: Inline error isolation, auto-draft saving, zero whole-batch wipes.

The bank credit & risk officer

• Environment: Managing corporate loan syndications across state and commercial banks (Mandiri, BRI, BCA).
• Mental Model: Cannot release credit disbursements until official AHU fidusia certificates are legally stamped.
• Behavioral Vulnerability: Rejects certificates if syndicated creditor proportions or subsidiary tax IDs are truncated.

Core Need: Guaranteed preferential rights, syndicated disclosures, audit-proof data.

The Ditjen AHU desk verifier

• Environment: Reviewing thousands of incoming legal transactions per hour with strict ministerial SLAs.
• Mental Model: Bound by state auditing standards. Cannot approve records with unresolved vehicle database mismatches.
• Behavioral Vulnerability: Overwhelmed by support tickets from notaries asking to manually unlock stalled billing vouchers.

Core Need: Pre-validated clean input, automated PNBP billing sync, decoupled revision queues.

COMPARATIVE ANALYSIS & ARCHITECTURAL FORK

International benchmarking: how modern registries handle revisions

Rather than guessing user behavior in isolation, we benchmarked how established collateral jurisdictions—such as Singapore's ACRA (BizFile+) and the UK Companies House Mortgage Registry—decouple typographical rectifications from formal debt variations.

Singapore (ACRA BizFile+)

Correction handling
ACRA provides a Notice of Error route for eligible clerical and typographical errors in filed documents. The correction route depends on the filing and error type.

Priority and security rights
Useful design benchmark: distinguish error correction from a substantive change. No blanket guarantee of charge priority is inferred from this guidance.

ACRA · Filing a Notice of Error

UK Companies House

Correction handling
MR04 records satisfaction in full or in part of a charge; it is not a general clerical-correction form. Companies House guidance describes court-ordered rectification under section 859M for relevant charge information.

Priority and security rights
Correction and satisfaction are distinct legal actions. The design should make the required route explicit rather than imply that every correction is administrative.

Companies House · Rectification of charges

Legacy Ditjen AHU (Indonesia)

Correction handling
Monolithic Failure: Any typo on a batch required full cancellation and re-entry from scratch.

Priority and security rights
Project-reported risk: cancellation and re-entry increased the risk of missing the 30-day filing deadline under Article 4 of PP No. 21/2015.

PP No. 21/2015 · Registration deadline

Our redesign (Option B)

Correction handling
State-Gated Recovery: Isolated clerical rectification (Rule A) from financial valuation changes (Rule B).

Priority and security rights
Original execution date and deed number remain immutable in the proposed workflow, maintaining the audit baseline. This is a product safeguard, not a guarantee of legal priority.

Option A

Easy Complexity

Incremental Form Polish: Keep the legacy linear wizard, add 60-second auto-save, and write clearer field error tooltips.

+ 3-week engineering turnaround.

+ Zero changes to existing ministerial database schemas.

- Fails to stop whole-batch failures on 50-item vehicle lists.

- Doesn't protect against the 30-day statutory expiration risk.

Verdict: Rejected • Too superficial for a legal crisis

Option B (selected)

Medium Complexity

Asynchronous Staged Validation & The "Perbaikan" Loop: Decouple entity identity, multi-creditor selection, and collateral tables into independent staging states with an inline revision loop.

+ Eliminates complete session discard on partial validation failure.

+ Directly safeguards notarial deeds from 30-day statutory expiry.

+ Accommodates multi-creditor syndications natively.

- Requires staged polling and new database draft models.

Verdict: Approved • Optimal balance of legal safety and engineering velocity

Option C

High Complexity

Cross-Ministerial Headless Auto-Verification: Full background sync between AHU, Korlantas (vehicles), ATR/BPN (land), and Dukcapil for automatic real-time pre-fill.

+ Completely frictionless zero-typing experience.

- Multi-agency intergovernmental MoUs required.

- Estimated rollout horizon exceeding 18-24 months.

- Any external API outage would halt all national credit registrations.

Verdict: Shelved • Unrealistic timeline and extreme dependencies

WHAT WE GOT WRONG

The unguided Excel bulk upload trap

Our initial prototype relied on a generic "Upload Excel / CSV" button to save notaries from manual typing. In practice, it caused widespread database exceptions.

Initial assumption vs. real-world breakdown

My initial assumption was that allowing clerks to upload spreadsheet files would instantly solve the data entry bottleneck for large corporate fleets. I was wrong.

In practice, notary clerks copy-pasted raw data sent from local car dealerships and regional banks. These spreadsheets were filled with invisible formatting chaos: rogue trailing spaces, hyphens in 17-character chassis numbers, and lowercase 'o' typed instead of numerical zero '0'. Because the legacy parser processed the file monolithically, the backend threw raw database SQL syntax errors. Notary clerks panicked, believing their accounts were suspended or the portal was hacked.

The Iterative Fix: We replaced blind file uploads with Guided Inline Linting & Data Sanitation. The interface sanitizes clipboard text in real-time, strips non-alphanumeric characters from VINs, visually highlights exact formatting errors in yellow, and allows inline corrections before committing rows to the database ledger.

PROCESS TRANSFORMATION

Macro flow: legacy all-or-nothing vs. staged recovery loop

Contrasting the 7-step legacy failure breakdown against the resilient staged intake protocol engineered in Figma, collapsing filing turnaround from weeks to under 15 minutes.

Old experience

Legacy Flow: 7-Step Friction (58.4% Mid-Funnel Abandonment • 30-day Legal Expiry Risk)

Start

→

Step 1

Deed data entry

Notary number & date

→

Typo Risk

50-item typing

Manual VIN entry

→

Crash

Single typo (#42)

Validation fails

→

Data Lost

Batch wiped out

Session timeout & reset

→

Billing Lock

Voucher expiry

SIMPONI unpaid

→

Time Out

30-day limit hits

Calendar window closes

→

Deadline missed

Swipe or scroll horizontally to follow the flow.

New experience

New Experience: Staged Verification & Modular Recovery (< 15 Mins Total)

Start

→

1. Request

Proses request

Unlock without deed reset

→

2. Multi-Bank

Syndicated creditors

Dynamic partner append

→

3. Obyek Check

Duplicate guard

Prevent double-pledge

→

4. Gating

Voucher tier lock

Step 4 gated by fee tier

→

5. Upload

Legal evidence

Akta perubahan bind

→

< 15m Certified

Swipe or scroll horizontally to follow the flow.

ARCHITECTURE & INTERACTIVE PROTOTYPE

The 6-step recovery engine: `Perbaikan Perubahan Sertifikat Fidusia`

Two revision policies separate descriptive corrections from valuation changes. The six stages below show the recovery architecture and its state boundaries.

Mode A · descriptive revision

Typographical corrections are editable; Step 4 (Nilai Penjaminan) remains locked to preserve the original valuation and voucher tier.

Mode B · valuation change

Debt valuation is active. The input is calibrated to the purchased PNBP voucher tier; descriptive stages remain locked.

Step 1: Multi-creditor binding (syndication)

Editable Canvas

Supports syndicated banking consortia. Users can append multiple lending partners (e.g. Bank Mandiri and BCA) with dynamic proportional quotas without exceeding table row bounds.

Step 2: Underlying agreement (perjanjian pokok)

Editable Canvas

Supports multi-currency debt facilities (e.g. IDR credit tranches alongside USD machinery facilities) without requiring distinct registration documents.

Step 3: Uraian Obyek Jaminan (vehicles & livestock)

Guarded Canvas

Structures vehicles (Yamaha NMAX with BPKB verification) and agricultural livestock (2 Ekor Sapi) with preserved physical photo evidence records.

Perhatian: Harap lakukan pengecekan melalui Pencarian Data Jaminan Fidusia untuk menghindari obyek ganda (double-pledging fraud).

Step 4: Nilai Penjaminan (voucher gating engine)

Locked State

Mode A Active: This step is disabled because descriptive revisions (typos, addresses, chassis numbers) do not alter financial debt valuation. This preserves state tariff integrity and avoids re-purchasing SIMPONI vouchers.

Step 5: Akta Notaris (audit stamp)

Read-Only Baseline

Immutable original notarial deed number and execution date preserved as audit baseline, preventing back-dating or fraud.

Step 6: Perubahan diff & confirmation

Verification Ready

Side-by-side comparison showing old baseline vs newly amended rows. Verifiers can approve revisions in seconds without full file re-audits.

KEY INTERACTION DECISIONS

Small decisions that prevented systemic failures

Three concrete UX interventions that stopped data entry errors from cascading into legal liabilities.

1. stepper re-ordering: moving "Obyek" before "Nilai"

The Problem: In the legacy system, clerks entered Nilai Penjaminan (Step 2) before declaring the actual asset list (Step 3). When clerks subsequently registered 50 fleet trucks with total market values exceeding the initial estimate, state revenue tariffs conflicted, forcing complete draft cancellations.

The Solution: We re-sequenced Step 3 (Obyek) to precede Step 4 (Nilai). The UI auto-tallies asset values first, ensuring the correct PNBP voucher tier is selected before committing legal covenants.

2. CTA hierarchy: "Simpan baris" vs "Kirim pengajuan"

The Problem: The legacy intake had a single ambiguous Kirim / Submit button at the bottom of the table. Clicking it pushed all 50 items simultaneously to external government clearing databases—and if item #42 failed, all 50 were discarded.

The Solution: We introduced a dual-layer CTA structure: row-level Simpan Sementara commits individual assets to local staging, while global Kirim Pengajuan is only enabled once 100% of rows pass syntax checks.

3. resolving 'o' vs '0' characters in VINs

The Problem: Dealership spreadsheets frequently contain the letter 'O' mistyped instead of the numeral zero '0' inside 17-character chassis numbers, along with hidden Unicode non-breaking spaces (U+00A0). Backend SQL parsers crashed on these invisible characters.

The Solution: We implemented client-side regex sanitization that executes during clipboard paste—auto-converting 'O' to '0' and stripping non-ASCII characters, eliminating formatting crashes before hitting the database.

DEFENSIVE DESIGN

Defensive design for public sector realities

Handling moments when external infrastructure fails to ensure notaries never lose progress.

If internet drops mid-entry (50-item fleet)

Browser IndexedDB Draft Persistence: When entering long vehicle portfolios, inputs are cached in local encrypted browser storage every 15 seconds. If the office connection drops or the browser crashes, 100% of entered serials restore upon re-opening.

If external police / vehicle database times out

Non-Blocking "Dalam Antrian" State: If the Korlantas external verification API hangs, the platform queues the unverified asset rows in the background rather than freezing the notary's screen, letting them complete remaining steps safely.

If multi-bank syndication split does not equal 100%

Real-Time Proportion Arithmetic Linter: When distributing loan collateral across syndicated banks, the system runs live percentage balance checks, warning users immediately if total allocations deviate from 100% before legal deed lock.

If SIMPONI billing voucher expires before payment

1-Click Voucher Regeneration: If a bank's accounting team takes too long to pay the government PNBP billing code, clerks can regenerate a fresh voucher code with one click without having to rebuild the verified registration application.

RESULTS

Measurable conversion & operational outcomes

Project-reported outcomes following rollout, based on the internal measurement methodology described below.

0

Reported deadline-related losses

Zero notarial deeds forfeited due to system restarts within the mandatory 30-day window.

-35%

Application drop-offs

Direct reduction in mid-funnel corporate filing abandonment at the asset collateral stage.

< 15m

Revision turnaround

Slashed from 3–5 working days to under 15 minutes via inline recovery drawers.

-65%

Helpdesk escalations

Drop in ministerial support tickets regarding locked records and expired billing vouchers.

Safeguarding national revenue & banking rails

Ditjen AHU processes over 12.2 million legal requests each year, generating more than $70M+ (IDR 1.1T) in annual non-tax state revenue (PNBP). By unblocking registration chokepoints, the redesign ensured uninterrupted credit flow for over $26B+ (IDR 400T+) in commercial banking and leasing assets nationwide.

Official Portal: Ditjen Administrasi Hukum Umum (Kemenkumham RI)

Estimated $5M+ (IDR 78B+) in re-issuance cost avoidance

By eliminating statutory deed expirations, notary associations and financial institutions avoid estimated tens of billions of Rupiah in re-drafting fees, repeated debtor signatures, and commercial dispute litigation expenses each year.

REFLECTION

The principle I took from this project

Design principle

"In enterprise and public-sector software, application drop-offs are rarely due to a lack of user motivation—applicants have a legal mandate to finish. When users abandon a government platform, it is almost always because the system forced an impossible all-or-nothing threshold. Decoupling complex legal validations into resilient, staged interactions transforms administrative friction into bulletproof compliance."

TEAM & COLLABORATION DYNAMICS

Core team & working relationships

Enterprise public infrastructure cannot be designed by committee. I owned the end-to-end architecture of this recovery engine as lead designer, working directly with executive leadership, product, and engineering.

Executive strategy & business sponsorship

Fariz Iskandar (CEO) & Danny Sabhara (Co-Founder / Director)

Fariz

Danny

Direct executive sparring partners on statutory compliance exposure, revenue protection guarantees with Ditjen AHU directors, and enterprise delivery milestone scoping.

Design ownership (this workflow)

Rajesh Siburian (Lead Product Designer)

LinkedIn

Portfolio ↗

Autonomous design ownership of the 6-step recovery state machine, non-destructive branching logic, and client-side VIN regex linter featured in this study. Note: Senior peers Yayang Rahmadina & Jordan Alfarishy covered parallel modules (New Registrations & Roya/Erasure) across the wider initiative.

Product management & delivery

Yusuf (Product Manager)

Spearheaded ministerial SLA alignment, sprint scope management, and dependency tracking with the Ministry of Finance's SIMPONI billing clearinghouse.

Frontend engineering & architecture

Anthony William Chandra (Frontend Developer)

Anthony

Engineering execution of the asynchronous staging state machine, client-side regex linting interceptors, and non-blocking voucher polling without requiring database schema refactoring.

APPENDIX • DATA SOURCES & VERIFICATION METHODOLOGY

Data sources & verification methodology

To maintain enterprise design rigor, every regulatory statute, economic metric, and telemetry datapoint in this study is grounded in traceable public documentation and internal production logs.

30-day statutory expiration registration deadline

Under Indonesian law, uncompleted fiduciary registrations past 30 days from deed execution exceed the statutory filing deadline; the consequences require case-specific legal review.

BPK RI: PP No. 21/2015, Art. 4 ↗

$70M+ (IDR 1.1T) annual PNBP state non-tax revenue intake

Official government performance reporting (LAKIP / LKjIP) documenting Ditjen AHU's annual revenue collections, where Fiduciary Collateral represents the primary transactional volume.

Official Portal: Ditjen AHU (Kemenkumham) ↗

$26B+ (IDR 400T+) credit base national collateral portfolio

Aggregate annual market volume of commercial motor vehicle financing, heavy machinery leasing, and business credit tranches registered under fiduciary collateral cessions across Indonesian banks and multifinance firms.

OJK: Financial Services Authority Statistics ↗

-35% drop-off & < 15m SLA production workflow metrics

Aggregated from 2,947 active notary filing sessions, system database audit tables, and Ditjen AHU helpdesk ticket resolution logs pre- vs post-implementation of the 6-step recovery engine.

Internal Telemetry (Lacak.io × Ditjen AHU)

$5M+ (IDR 78B+) cost avoidance legal re-issuance savings

Calculated based on standard statutory notarial deed re-drafting fees under Government Regulation (PP) No. 21/2015 multiplied by the volume of filings rescued from statutory deed invalidation.

BPK RI: Government Regulation No. 21/2015 ↗

UI architecture & state models design artifacts

Production design workspace detailing the 6-step revision stepper, voucher gating state models, multi-entity identity cards, and document attachment modals.

Production Design Workspace (Ditjen AHU / Lacak.io)


Outcome note: Project results are reported by the project author. National financing and revenue figures describe broader context, rather than impact attributable to this redesign. Cost avoidance is an estimate.

Frequently Asked Questions

Government portals are notoriously rigid and politically resistant to change. How did you get Ditjen AHU and technical vendors to approve moving from monolithic validation to an asynchronous recovery engine?

Why did you engineer strict voucher-tier gating (Rule A vs Rule B) instead of allowing users to edit any field freely during a revision?

Across the national registry ecosystem, what did you personally own, and how did you verify that your solution actually worked?

Share

Text of this case study
Presentation slides
Case study PDF

JEZ

Explore the decisions and systems behind meaningful products.

Create a free website with Framer, the website builder loved by startups, designers and agencies.