FLIK Strategy & Roadmap

INT-12-MCA-Credit-Engine

7 min read

đź“„ Initiative Spec: INT-12 MCA Credit Engine (LOS)

Status: 🔵 Active (Definition Stage) Owner: azmi Strategic Intent: Build a paperless, automated high-velocity lending engine to dominate the D2C financing market. Automate data collection, processing, and automate financial reports all with the least amount of coding

đź“Š Executive Summary

Situation: FLIK is scaling its Merchant Cash Advance (MCA) product to bridge the credit gap for Indonesian D2C brands. While we control the repayment flow via our payment infrastructure, the front-end “Origination” process remains a manual, high-friction operation relying on WhatsApp document dumps and manual Excel scoring.

Complication: The manual process creates significant operational bottlenecks and regulatory risks:

  1. Compliance Risk: SLIK reporting has a strict 5-day window for file submission; manual handling risks suspension.
  2. Credit Blindspots: Current snapshot-based scoring misses “Shadow Debt” (non-bank P2P loans) and revenue volatility, leading to potential under-pricing of risk.
  3. Scalability Wall: The current high-touch model cannot support the target volume of Rp 6B+ without a linear increase in headcount.

Strategic Solution: Deploy a 2-phase MCA Credit Engine (LOS):

  • Phase 1 (The Brain): Automates data ingestion, SLIK compliance bots, and a 5-factor scoring model that captures hidden debt and marketing efficiency.
  • Phase 2 (The Scale): Integrates Privy for e-signatures and Asli RI for automated KYB/Authority validation, removing all manual paper/PDF steps.

Expected ROI:

  • Velocity: Reduce Time-to-Disburse from 3 days to <12 hours.
  • Efficiency: 10x throughput capacity without increasing Ops/Finance headcount.
  • Risk: 20% reduction in potential default rate through better visibility into shadow liabilities.

1. The “Why” (Strategic Gap Analysis)

1.1 Current State (Manual Friction)

  • Data Ingestion: Sales collects PDFs/Screenshots via WhatsApp & Email.
  • SLIK Compliance: Manual inquiry followed by manual file renaming and zipping to meet the 5-day OJK rule. High risk of suspension.
  • Risk Blindspots: Scoring relies on SLIK, missing “Shadow Debt” from P2P/Platform lending.
  • Legal Execution: Manual PDF generation and Privy preview links.
  • Disbursement: Manual configuration of repayment logic by engineers per loan.

1.2 Target State (Automated Engine)

  • Data Ingestion: Unified Merchant Portal for self-service upload.
  • SLIK Compliance: Automated Bot that handles Inquiry + Zipping + Naming + Reporting instantly.
  • Risk Visibility: Calculator v2 integrates Shadow Debt and Marketing Performance variables.
  • Legal Execution: Zero-touch Privy API + Asli RI for authority validation.
  • Disbursement: Automated orchestration triggered by “Green Signal” from internal approvals.

1.3 The Gap (What we are building)

We are bridging the gap between a “High-Touch Consultancy” model and a “Fintech Infrastructure” model. The priority is to remove human dependency from the “Compliance” and “Scoring” layers first.

Strategic Upgrade: Scoring Model Evolution

  • Current Model (“Checker”): Manual Excel snapshot. Asks “Can they pay us back today?” Flawed by ignoring revenue variance and hidden platform debt.
  • New Model v2 (“Risk Engine”): Algorithmic weighted score. Asks “Will they survive the next 6 months?”
    • Upgrade 1: Stability over Average. Punishes spiky revenue (High CV) even if the average is good.
    • Upgrade 2: Shadow Debt Visibility. Explicitly subtracts Shopee/Tokopedia loans.
    • Upgrade 3: Growth Quality. Distinguishes between “Bought Revenue” (Ads) vs. “Earned Revenue” (Retention).

Comparison: Current vs Target Logic

FeatureOption A: Current Toolkit (Manual)Option B: Automated Risk Engine (Target)Why Upgrade?
Revenue HaircutStatic 15% discount on Average.Lowest Month Baseline.Uses the “Worst Case” month as the lending base. Simple & Safe.
Debt LoadChecks SLIK (Bank) only.SLIK + Shadow Debt (Platform/P2P).(Lowest_Rev - Shadow_Debt) = Safe_Base. Prevents over-leveraging.
CollectionAutomated (Revenue Split via FLIK).Automated (Revenue Split via FLIK).Core mechanism is already automated. Focus shifts to streamlining security documentation.
ScoringLinear Sum.Weighted 5-Factor Model.More accurate risk prediction.

MCA Process Flow

2. Process Flow Alignment (BRD)

Based on the Master Flow diagram, the end-to-end MCA lifecycle is categorized into four distinct functional blocks:

A. Initialization & Application

  1. Lead Capture: Sales provides the application link to the Merchant.
  2. Consent (Legal Bridge): Merchant provides digital consent for data processing (Checkbox).
  3. Identifier Check: System classifies the entity (PT, CV, Individual).

B. Verification & Data Enrichment

  1. Commercial Ingestion: Automated upload of Marketplace PDFs and Bank Statements.
  2. Site Visit: System-captured photos with metadata verification .
  3. SLIK API Integration: Real-time inquiry and automated compliance reporting (Zipping/Naming).

C. Credit Scoring & Approval

  1. Scoring Engine (Calculator v2): Calculation of risk score including “Shadow Debt” (platform lending not in SLIK).
  2. Commercial Memo: Auto-generation of the investment brief.
  3. Decision Matrix: Tiered approval flow (Manager → Finance → CEO for high-value).

D. Legal Execution & Capital Deployment

  1. Legal Sign-off (Phase 2): Privy e-signature and Asli RI authority validation.
  2. Disbursement Orchestration: Auto-triggering of funds via payment gateway (Xendit) once internal approval signals are green.
  3. Project Configuration: Initializing the “Bulk Purchase” object with specific recollection logic (Daily/Weekly).

2. The “What” (Solution Components)

2.1 Phase 1: The Decision Engine (Automated Credit Assessment)

Focus: create a 1 touch application that will require little manual analyst review and quick approval

  • Commercial Data Portal:

  • Centralized Data Collection:

    • Gap: WhatsApp dumps and fragmented email threads.
    • Fix: Secure Merchant Portal for direct upload of Marketplace PDFs, Bank Statements, and Legal Docs.

  • Data Processing & Extraction:

    • Gap: Manual data entry into Excel calculators.
    • Fix: Automate the process by implementing it in 3 tier step.
  1. Data Audit for core and booster Metrics to ID required value to perform a more holistic extraction of the dataset.
  2. Identify the type of data to ensure what it is for. To know how to process and validate.
  3. Test with a variety of sample that is validated to ensure that there is no flaw or what has to be improved.
  • SLIK Automation Bot:
    • Gap: Manual zipping/renaming risk.
    • Fix: API-driven inquiry with auto-compliance (Zip + Consent ID).
  • Scoring Calculator v2 (5-Factor Model):
    • Gap: Missing platform lending debt (Shadow Debt) and limited scoring variables.
    • Fix: Algorithmic scoring integrating GMV Stability, Quality, Shadow Debt, and Marketing Efficiency.
  • Automated Finance Reporting (Loanbook & Risk Dashboard):
    • Gap: Manual Finance Excel tracker.
    • Fix: Live dashboard with NPL, Principal Outstanding, and Concentration risk.

2.2 Phase 2: The Paperless Scale (Closing the Trust Gap)

Focus: Solving the “Scale” - Legal Execution and Identity.

  • Privy API Integration:
    • Gap: Manual PDF previews sent via email.
    • Fix: End-to-end API signature request triggered by approval.
  • Asli RI Validation:
    • Gap: Manual visual verification of Director’s KTP vs. Akta.
    • Fix: Automated API check to validate Director authority and corporate data.
  • Approval Dashboard:
    • Gap: Slack-based approvals.
    • Fix: Centralized “Checkbox” dashboard for Manager/CEO sign-off.

2.3 System Control & Governance (Phase 2 Configuration)

Focus: Management levers to control risk and delegation.

  • Configurable Operating Modes:
    • Mechanism: Central “Risk Switch” to toggle the engine’s aggression levels.
    • Default Parameter Sets:
      ParameterConservativeStandardHigh-Growth
      NPL Trigger>1.5%1.0% - 1.5%<1.0%
      Max Advance Multiple0.8x1.2x1.5x
      Max Tenor90 days120 days180 days
      Auto-Approval Limit200M500M750M
    • Feature: System recommends mode based on Portfolio NPL and Funding Headroom.
  • Dynamic Authority Matrix:
    • Mechanism: Rules engine to route Approval Memos. | Level | Max Amount | Min Score | Conditions | | :--- | :--- | :--- | :--- | | Auto | ≤ 500M | ≥ 80 | All within policy | | Analyst | ≤ 300M | ≥ 65 | Standard only | | Manager | ≤ 1B | ≥ 50 | Minor exceptions | | CEO | > 2B | Any | Large/High Exception |
    • Feature: All routing thresholds are configurable via the Admin Panel.

3. Implementation Phases

Phase 1: The Decision Engine (Core Data & Analysis)

  1. Central Data Collection: Portal for ingestion of Marketplace PDFs, Food Delivery data, and Bank Statements.
  2. Merchant Identifier: Automated classification of entity type (PT, CV, Individual).
  3. SLIK Automation (Analysis): Inquiry bot with automated zipping/naming compliance (5-day max requirement).
  4. Scoring Input (Calculator v2): Integration of Shadow Debt (Platform Lending) and Marketing variables into the scoring model.
  5. Memo & Orchestration: Automated Commercial Memo generation and disbursement orchestration logic.

Phase 2: The Paperless Scale (Legal & Trust)

  1. Privy API Integration: Automated e-signatures for agreements.
  2. Asli RI Validation: Corporate authority and KYC verification.
  3. Approval Dashboard: Centralized checkbox-based management workflow.

4. Key Metrics (ROI)

  • Time-to-Disburse: Target < 48 hours from application.
  • Operational Headcount: Ability to handle 100x volume without increasing the Finance/Ops team size.
  • Default Rate: Reduced via better visibility into platform-specific debt.

5. Action Items

Action Items for Phase 1: Automate and create the most valuable info and a clear and compliant report for analysis. This will allow automated key values with more insight to our customer with a higher verification to ensure less fault or error

  • Create the Landing page (application done)
  • Formalize SLIK zipping/naming logic with the Engineering team (Critical for compliance)
  • Have the model audit with legal to review if all process is following procedure
  • Train and improve to remove manual entry from the model by making it less than 50 min for it to get processed

Graph View