0% found this document useful (0 votes)
6 views24 pages

6estates Introduction

6Estates specializes in providing end-to-end AI solutions for financial services, enhancing growth and efficiency through business-ready Generative AI. The company serves over 30 enterprises, focusing on lending, risk assessment, and operations, with a strong emphasis on accuracy and security. Their offerings include customized Large Language Models, Intelligent Document Processing, and automated workflows, delivering significant operational cost reductions and improved accuracy in document handling.

Uploaded by

williamsidharta
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
6 views24 pages

6estates Introduction

6Estates specializes in providing end-to-end AI solutions for financial services, enhancing growth and efficiency through business-ready Generative AI. The company serves over 30 enterprises, focusing on lending, risk assessment, and operations, with a strong emphasis on accuracy and security. Their offerings include customized Large Language Models, Intelligent Document Processing, and automated workflows, delivering significant operational cost reductions and improved accuracy in document handling.

Uploaded by

williamsidharta
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

Driving enterprise

growth and efficiency


with Generative AI
July 2025
Spun-off in 2014 from the NExT AI Research
Centre, by NUS and Tsinghua University

Team of AI specialists supported by proven


investors and entrepreneurs, including the
Built for Lenders, trusted Singapore government

by Leaders — 6Estates
Enhancing growth and efficiency for financial
delivers end-to-end AI services through business-ready GenAI
transformation for
lending, risk, and Serving 30+ enterprises across the region,
including leading banks, NBFIs, logistics and
operations retail companies

ISO 27001 certified, reflecting commitment to


information security management
TRACTION
We already serve 30+ leading enterprises across the region,
particularly across Financial Services and Logistics

Traditional
Banks
At a Glance
Digital Banks ■14 Customers in
Indonesia

Non-Bank ■17 Customers


Lenders from other
countries/region
(Singapore,
Rural Banks Malaysia, Hong
Kong, China)

■Majority of
Other FSIs customers in are
in Financial
Services
Non-FSIs
Overview: 6Estates’ Enterprise AI Solutions,
Powered by Large Language Models

Customised LLMs and Intelligent Document Workflow Automation Financial Statement LLM
Applications Processing (SaaS) for Lenders (SaaS) Agent (SaaS)

LLM IDP FAAS FSAgent


■ End to end support for Large ■ Comprehensive AI-powered ■ Automates the entire loan ■ Automates in-depth analysis of
Language Model (LLM) document processing and process, from onboarding, fraud financial statements
customization and deployment analysis platform assessments and risk scoring to ■ Capable of document analysis and
(including hardware and base ■ Supports all industries, document credit analysis and monitoring Q&A, extraction of critical data,
models, fine tuning, and app types, formats and languages ■ Meeting a variety of lending use reasoning from long-form
development) with >99%+ accuracy cases, including SME, Commercial, documents, precise calculations
Auto and Consumer finance and input selection, and
trend-based forecasting

6Estates’ Value Proposition

✓ Highly flexible and adaptable AI solutions, optimized to be business-ready and user-friendly


✓ Drives significant reductions in operational costs and human errors
✓ Close to 100% output accuracy by integration of AI-assisted HITL function
✓ Domain expertise in financial services and logistics/trade
Intelligent Document
Processing (IDP)
Platform
Document Automation for All Industries
6Estates offers industry-agnostic and
domain-specific document processing solutions

IDP for Lending (FAAS) IDP for Logistics & Trade

Analytics & Analytics &


Applications Insights Insights
Domain-specific Solutions … more to come
built on IDP
Domain Specific Domain Specific
Rules Checks Rules Checks

File Detection Information Cross Doc General


Digitalization
& Auto Split Extraction Matching Rules Checks
Platform
Intelligent Document
Processing (IDP) platform • Bank Statement • Credit Note • Annual Return any semi-structured
• Credit Bureau Report • Financial Statement • Invoice/PO/DO & unstructured
• ACRA Bizfile • Company Deed • Medical Bill ……. documents

6Estates Domain-Specific Large Language Model (LLM)


Model
Domain-Specific Large
Partner & Open Source Base Model Second Pre-training + Supervised
Language Model Underpins + 6Estates Domain-Specific Fine-tuning + Retrieval
Reinforcement Learning from
Platform and Applications Human Feedback (RLHF)
Foundation Model and Dataset Augmentation
6Estates' IDP capabilities

Digitization and Extraction File Detection & Auto Split


Turn fields, tables and text from unstructured Upload multiple documents at once
documents into structured data (accepts all image formats and even ZIP files)

Rules Checks Cross Document Matching


Run automated rules checks and verification
Compare and verify data from multiple sources
(e.g. trade rules and standards checks)

Semantic Search Analytics & Insights


‘Talk’ to the results to easily find answers Calculate metrics and output reports
Current IDP Document Coverage

Onboarding
Trade Related Financial Collateral
Documents for
Documents Documents Documents
Legality and Identity

● ACRA Bizfile, ● Airway Bill, ● Bank Statement, ● Buku Pemilik Kendaraan


● Company Deed*, ● Bill of Lading (BL), ● Cheques, Bermotor (BPKB),
● Singapore NRIC, ● Bill of Exchange (BE), ● Credit Note, ● Surat Tanda Nomor
● Hong Kong Annual Return ● Delivery Order, ● Credit Bureau Singapore, Kendaraan (STNK)
● e-KTP, ● Document Export, ● Debit Note,
● ID Card, ● Invoice, ● Financial Statement*,
● Kartu Keluarga, ● MT700, ● Hong Kong Annual Return,
● NPWP, ● Packaging List, ● Receipt,
● NIB*, ● Purchase Order, ● Payslip, etc
● Passport, etc ● Express Waybill
● Schedule Remittance or
Letter, etc

* Being Refined
IDP supports a variety of business use cases

Selected use cases and document types

Loan Assessment &


Customer Onboarding and Trade Finance
Monitoring
KYC/KYB • Letters of Credit
• Bank Statements
• Identity Documents • Commercial Invoices
• Financial Statements
• Company Deeds • Bills of Lading
• Credit Bureau Reports
• Company Filings… • Packing Lists…
• Collateral Documents…

Procurement & Financial


Insurance Claims Sales Processes Reconciliation
• Invoices and Receipts • Inquiries and Quotations • Invoices
• Ownership Documents • Purchase Orders • Credit / Debit Notes
• Policy Documents… • Delivery Orders… • AP / AR
• Bank Statements…
MRC delivers significant benefits over
manual processes and OCR technologies

Easily configurable

Accuracy up to 99%+, supported by AI and HITL

Ability to adapt and generalize to new formats

Compatible with complex documents and tables

Accuracy improvement over time

Insight and report generation

High document processing speed

Seamless integration into existing systems


Zero Learning Extraction

6Estates' AI can read unknown (never trained) documents

CONFIGURE
Instead of configuring the x-y
longitudinal (position) of each field
(like OCR), users only indicate fields
they need with natural language

UPLOAD
A document type only needs to have
field(s) configured once.
Al will capture required data from
further documents, in spite of
differing formats
Summary of Live Use Cases (pg 1/2)

USE CASES USERS WORKFLOWS DOCUMENT TYPES COUNTRIES


▪ SME banking ▪ Data extraction from financial and ▪ Bank statements
identity documents
▪ Commercial banking ▪ Financial statements ▪ Indonesia
▪ Transaction fraud and risk detection
Loan application ▪ Rural banking ▪ Credit bureau report ▪ Singapore
▪ Consolidation and restatement
assessment and loan ▪ Digital banks ▪ Ownership documents ▪ Malaysia
▪ Credit analysis and rules checking
monitoring (FAAS) ▪ Auto finance ▪ Payslip ▪ Hong Kong
▪ Risk variables and indicators
▪ Mortgage finance ▪ Tax documents
▪ Reporting to feed into banks’ risk
▪ Credit card model ▪ Identity documents…
▪ ID (KTP/NPWP/BPKB/KK)
▪ Traditional Banks
▪ Data extraction from identity ▪ Singapore NRIC
▪ Digital banks ▪ Indonesia
Onboarding and documents ▪ Passport
▪ Rural banks ▪ Singapore
KYC/KYB ▪ Cross-document checks to verify data ▪ ACRA bizfile
▪ Non-bank lenders consistency ▪ Annual return
▪ Payments
▪ Company deeds…
▪ Read email and extract inquiry needs
▪ Align with internal products and prices ▪ Inquiry Order
▪ Companies with ▪ Generate and send back the quotation Purchase Order
Front office – Sales ▪
email/mail-driven sales ▪ Read purchase order and connect with ▪ Singapore
process automation ▪ Delivery Order
processes SAP system
▪ Data from ERP
▪ Generate Delivery Orders via SAP
system
Summary of Live Use Cases (pg 2/2)

USE CASES USERS WORKFLOWS DOCUMENT TYPES COUNTRIES


▪ Extract Invoice information
▪ Extract Purchase Order information ▪ Invoice
Back office –
▪ Companies with heavy and match with Invoice ▪ Purchase Order ▪ Indonesia
Procurement &
procurement and ▪ Extract Delivery Order information ▪ Delivery Order ▪ Singapore
financial
finance works and match with Purchase Order ▪ AP/AR ▪ Thailand
reconciliation
▪ Reconciliation between AR and bank ▪ Banks statements
statements
▪ Invoices
▪ Motor Insurance ▪ Data extraction from evidence ▪ Receipts
Insurance claims ▪ Singapore
▪ Health Insurance documents and policy documents ▪ Ownership documents
processing ▪ Hong Kong
▪ P&C Insurance ▪ Data reconciliation ▪ Medical bills
▪ Insurance policy documents
▪ Companies with need to
▪ Financial documents e.g.
extract specific ▪ Indonesia
▪ Data extraction annual reports, loan
Financial Statement information from
▪ See FS Agent for enhanced financial agreements, earnings ▪ Singapore
Analysis long-form documents,
statement analysis transcripts, mortgage ▪ Malaysia
such as investment
agreements, etc
companies
FAAS:
Finance-as-a-Service
Lender Workflow Automation
FAAS drives
faster, more 70x 80%
efficient growth Faster
processing
Less errors vs.
manual
for lenders

4
Automates lending processes
from end-to-end

Designed for the financial


services industry, from
76% Layers of security
traditional banks and SME Cost savings and privacy
lenders to rural banks and protections
non-bank financiers
FAAS delivers clear business value, on top
of operational efficiency
Credit Score Determining credit health score based on
comprehensive analysis of bank statements
Manipulated Bank Statement Indicator,
Fraud Score
Fraudulence Bank Statement Indicator, etc. Credit analytics &
Using by-client recurring income and by-vendor
scoring based on 37
Payment Ability Analysis factors
Analytics payment data
Application & Insights Generate F/S compiling bank statements and
Financial statement recasting
FAAS on compare with submitted F/S, etc.

IDP
Anti-Fraud Check Account Consolidation
Check 100+ business rules.
Automatically generate 114
Domain Authentication Cross-doc consistency checks data fields related to bank
Specific statements
Rules Checks Entity Relations Smart categorization

File Detection Information Cross Doc General


Digitalization
& Auto Split Extraction Matching Rules Checks
Platform
Intelligent Document
Processing (IDP) platform …… Extract 30+ fields from bank statements from multiple banks
over multiple periods in different languages/currencies

Model
Domain-Specific Large Language Models Underpin Platform and Applications
FAAS combines human expertise with AI to
deliver business-ready results
FAAS is more than an efficiency tool – it is a business expansion tool designed for
bankers to serve and grow their SME customer base

1 AI PROCESSING 2 HUMAN VALIDATION 3 INSIGHT GENERATION

◆ AI classifies documents and extracts ◆ AI-assisted HITL examines and ◆ Platform categorizes and analyzes
data from lending documents (e.g. validates information identified by every transaction
bank statements, identity docs) IDP to require human intervention ◆ Outputs key metrics and risk
◆ AI organizes data into structured indicators for credit approval,
database pre-disbursement, loan monitoring

AI-driven IDP alone can achieve Addresses the last-mile problem to Capable of 50+ risk rules checks and
>85% accuracy enhance accuracy to > 99% 80+ credit variables

6Estates’ AI-driven process enables fast,


accurate and in-depth credit assessment
DOCUMENT INPUT FAAS PROCESSING FLOW AI GENERATED REPORT
1 2 5 8
Multiple bank statements Machine Reading Bank Statement Formatting,
Business Activity Indicators
in different quality and Comprehension (MRC) for Currency Standardization and
Information Extraction Consolidation ● Graphs of monthly and daily
image/ file types
fluctuation of income and
expenses
● Summary of cash balances,
3 6 cashflow, transaction activities

Automatic Rules Validation MRC for Transaction Categorization Transaction Activity Summary
Business vs. non-business transactions ● Breakdown of business and
● Format Checks
● Type of transaction by RTGS, non-business transactions
● Out of Bounds Dates
Clearing, or cash basis ● Monthly supplier-buyer
● Debit/Credit/Balance
● Circular Transactions activities
Checking
● Related Party Transfers
● Fraud Rules Checks: e.g. Re-Casted Income Statements
● Taxes
RTGS cannot occur on ● Yearly, quarterly, monthly
● Assets/Securities Sold ● Operational and credit metrics
Public Holiday, etc.
● Admin Fees
● Bills Risk and Fraud Indicators
● Payrolls ● Consolidated report of
4 ● High-amount transactions Manipulations, Fraud, and Risk
AI-assisted HITL Indicators (loan related)

7
Automatic rules-based transaction
insertion or deletion
Introduction to
FS Agent
Automate financial analysis with AI
FS Agent automates in-depth analysis of
financial statements
Delivers comprehensive insights, and reduces processing time from days to minutes

USER UPLOADS AND


1 CONFIGURES 2 FS AGENT ANALYZES 3 REPORT GENERATED

● User uploads financial ● FS Agent analyzes document(s) ● FS Agent outputs summarized


statement(s) in e-PDF or scanned using AI and LLM Agent findings and insights from
format (multiple languages technology documents and external sources
supported) [(via Word, Excel)]
● FS Agent can run quantitative
● User selects default standard and qualitative analysis, and ● Users can edit report content as
analysis, and/or configures draw information from external needed, or “interact” with the
custom analysis on our intuitive sources (e.g. news, 3rd party documents via a Q&A function
platform databases)
FS Agent: Detailed process

Start Analysis S1: Business Summary S4: Risk Factors S6: Financial Data Analysis

● Basic company background


● Ongoing litigation ● Balance sheet
● Main business activities
● Outstanding debts and ○ Standard financial fields
● Customer base
insolvency risks ○ Statement-specific
SELECT & CONFIGURE ● Target industries Auto-generated
● Unexpected / extraordinary non-standard financial
Standard and/or ● Material events Editable Memo
events fields
Custom Analysis ● Related parties
● Market risks ○ Domain-specific metrics
● Credit risk and margins, including ● S1 to S6
● Liquidity risk current ratio, W/C ratio, ○ Takeaways from
S2: Management Overview ● Other potential risks ROE, ROA, Debt-to-Equity, analysis by
etc. section
● Income statement ○ Output of
● Background on Directors and ○ Standard financial fields quantitative
E-pdf Scanned data (in table /
Key Executives ○ Statement-specific
S5: Shareholder Information
● Executive remuneration non-standard financial metric form)
UPLOAD ● Strategic priorities fields ● S7: Recent
Financial Statement(s) ● List of shareholders ○ Domain-specific metrics Company and
● Share issuances and margins Industry news
● Share buybacks ● Cash flow statement ● S8: Results of
S3: Management Perspectives ● Information on share buyback ○ Net cash flows from Custom Analysis
policy operating activities (CFO)
● Dividend policies and ○ Net cash flows from
● Views on financial condition activities investing activities (CFI)
● Views on changes in economic ● Summary of historical ○ Net cash flows from
background activities related to dividends financing activities (CFF)
● Views on company outlook and share capital ○ Beginning and ending
cash
Key benefits of FS Agent

✓ Comprehensive analytics capability: capable of robust


quantitative and qualitative analysis (pre-set analysis and Core FS Agent
customized)
Capabilities
✓ Document analysis and
✓ Powerful and flexible platform: able to analyze multiple Q&A functionality
documents together, even documents with differing formats
✓ Extraction of critical
and time periods
data and reasoning
from long-form
✓ Highly adaptable: able to gather relevant data, calculate documents
results, summarize information, and draw from external
resources without rigid rules or direction ✓ Precise calculations
and input selection
✓ Easily configurable: facilitates convenient free-style ✓ Trend-based
interaction with document content and configuration of forecasting
customized analysis (no coding required)
Customer pain points
‐ Lack of standardized, comprehensive and reliable
Use Cases: financial database for SMEs and private companies →
as a result, manual analysis of financial statements is
FS Agent ‐
critical
Tedious to generate memos based on large volumes of
long-form financial statements
‐ Existing tools are typically limited to standard,
non-configurable analysis
‐ Existing tools only support e-pdf, while non-standard
formats like scanned copies are not supported

6Estates’ solution
Investment and credit analysis ‐ FS agent: LLM-based AI tool for financial statement
analysis
‐ Memo/summary generation: AI auto-generation
‐ Customized configuration: clients can customize the flow
● Financial statement analysis involves trawling through and analysis based on their priorities
hundreds of pages and across multiple years, resulting ‐ Multi-year capability: Combine multiple years statements
in tedious and time-consuming work and provide a comprehensive analysis

● Users in the investment and trading industries can use


FS Agent for investment analysis
● Users in the lending industry can use FS Agent to
analyze business operations and credit quality for their Customer expected outcomes
current and potential commercial borrowers ‐ Comprehensive memo generated in minutes
‐ Easily configurable analysis and output format
‐ Consolidated analysis for documents from multiple
years
Thank you!

luanhuanbo@[Link]

[Link]

[Link]/company/6estates

You might also like