Case study · Housing finance · AI
Art Housing Finance Deployed Offline Aadhaar Masking in One Month at 99% Accuracy
Art Housing Finance is a home loan finance company in India. RBI guidelines require every Aadhaar number stored in its document repository, new or historical, to be masked.
Its lean IT team needed a solution it could deploy without cloud access, integrate with the existing document management system (DMS) and take live within one month, without a dedicated AI team or a six-month vendor evaluation.
- 99% masking accuracy
- 0.3 seconds per file
- Live within one month
Results at a glance
100%
RBI compliance for Aadhaar masking across the full document repository
99%
Masking accuracy across document types and formats
0.3 sec
Processing time per file, in line with daily intake
1 month
From scoping to live deployment, including DMS integration
The Challenge
Housing finance documents vary more than they first appear, which makes Aadhaar masking operationally complex. The same problems surface whether a lender processes hundreds of files a day or thousands.
Documents Span Formats That Behave Differently
Aadhaar cards, loan applications, CIBIL reports, and bank statements each present Aadhaar data differently, and PDFs, scanned images, and text files each need their own processing logic. A tool that is accurate on one format will miss Aadhaar numbers on another.
Daily Volumes Leave No Room for Bottlenecks
Loan files are generated, processed, and stored every day, and operations do not pause for compliance. Masking has to keep pace with daily intake without creating queues or delays in active workflows.
A One-Month Deadline Is Tight
Regulatory timelines are fixed. A solution that takes three months to configure, test, and deploy does not fit a one-month compliance window. Every component, from deployment to DMS integration, had to be delivery-ready from day one.
Data Cannot Leave Internal Infrastructure
A regulated housing finance company cannot send loan files and KYC documents to an external service for masking. Processing has to happen on internal servers, offline, with no external connectivity and a complete audit trail.
The DMS Cannot Be Disrupted
Branch teams rely on the document management system every day. A masking solution that forces workflow changes or causes downtime during implementation creates operational risk that outweighs the compliance benefit.
The Solution
HabileLabs designed and deployed an AI-powered Aadhaar masking and document processing system around Art Housing Finance's compliance requirements, infrastructure constraints, and integration needs.
Intelligent Document Processing
AI and ML-based classification identifies each financial document type and routes it before masking begins. Aadhaar cards, loan applications, CIBIL reports, and bank statements each get the right processing logic, with no manual routing.
AI-Powered Aadhaar Detection and Masking
OCR, computer vision, regex, and AI work together to detect Aadhaar numbers and mask sensitive data across all document types, handling variation in scan quality, layout, and format. Accuracy reached 99% across Art Housing Finance's full daily document volume.
High-Throughput Processing Engine
Each file is processed in 0.3 seconds, so masking keeps pace with daily intake without queues or backlogs in active operations.
Secure Offline Deployment with Docker
The system is containerized and runs on Art Housing Finance's internal Ubuntu server. It needs no internet access at any stage and required no changes to security policy or infrastructure configuration.
DMS Integration via RESTful APIs
RESTful APIs connect the masking solution directly to the existing document management system. Integration was completed within the one-month timeline, with no workflow changes or downtime for operational teams.
Business Impact
100% RBI Compliance Achieved
Aadhaar masking applied accurately across the full document repository, in full compliance with RBI guidelines.
99% Masking Accuracy at Scale
Consistent detection and masking accuracy across document types and formats, every day.
0.3 Seconds per File
Throughput kept pace with daily document intake, with no queuing delays.
Live Within One Month
Deployment, testing, and DMS integration completed within the one-month compliance window.
Offline, with No Infrastructure Changes
Runs entirely on the internal Ubuntu server with no internet access. Security policy unchanged.
No Operational Disruption
RESTful API integration with the existing DMS caused no downtime or workflow changes.
A shared challenge across housing finance and BFSI
What Housing Finance Companies and NBFCs Can Take from This
Art Housing Finance's situation is common. Every housing finance company, NBFC, or lending institution that has collected Aadhaar-based KYC over the past several years carries a version of this obligation. The trigger varies: an RBI inspection, an internal audit, or a branch expansion that surfaces legacy document liability. The underlying problem is the same: Aadhaar data sitting in document repositories with no automated masking in place.
Smaller institutions often face it with fewer resources: a lean IT team, no in-house AI capability, and a compliance window that leaves no room for vendor selection delays. The Art Housing Finance engagement was designed for those constraints: Docker-based offline deployment, a pre-built document classification layer, and a RESTful API that connects to the existing DMS, with no infrastructure rebuild, no cloud dependency, and no retraining required.
Housing Finance Companies Processing Loan Documents Daily
Daily volumes need masking that runs at processing speed, not batch speed. At 0.3 seconds per file, masking happens in line with intake rather than as a separate overnight process.
Institutions with a One-Month or Near-Term Deadline
The Art Housing Finance engagement went from scoping to live deployment, including DMS integration, within one month. When a deadline is close, a delivery-ready solution is the only workable path.
BFSI Organizations That Must Stay Offline
Docker-based deployment on an internal server with no internet connectivity is the standard model. It works in Ubuntu environments, restricted networks, and security-first IT setups without exceptions.
Organizations That Cannot Afford DMS Disruption
RESTful API integration plugs the masking solution into the existing document management system. Workflows do not change, and branch teams do not notice the transition.
Why Art Housing Finance Chose HabileLabs
Proven AI Capability in BFSI Document Processing
Art Housing Finance needed a partner that had already solved this class of problem in regulated financial environments, not one learning on its engagement. HabileLabs' track record in BFSI document compliance meant the solution was built to sector standards from the start.
Compliance and Accuracy Together
Art Housing Finance needed both 100% RBI compliance and 99% masking accuracy, and neither was treated as secondary. Both were delivered within the same one-month deployment window.
Working Within Existing Constraints
Offline Ubuntu server, no internet access, existing DMS, one-month timeline: every constraint was met without asking Art Housing Finance to change its infrastructure or extend its deadline.
“The requirement wasn't just accuracy, it was accuracy within an infrastructure we couldn't change and a timeline we couldn't move. HabileLabs delivered both.”
FAQs
Frequently Asked Questions
Yes. The solution processes each file in 0.3 seconds and is designed to match your daily document intake without creating queues or delays in active workflows.
Yes. The Art Housing Finance engagement went from scoping to live deployment, including DMS integration, within one month. A delivery-ready approach with defined components is what makes that timeline achievable.
Yes. The AI engine combines OCR, computer vision, and regex to handle variation in scan quality, document layout, and format. It delivered 99% accuracy across Art Housing Finance's full range of documents.
Yes. RESTful API integration connects the masking solution directly to your DMS. Your teams do not need to change their workflows, and there is no downtime during or after implementation.
Yes. The solution runs on your internal server via Docker and needs no internet connectivity at any stage. Art Housing Finance's deployment ran on an internal Ubuntu server with no external access.
Yes. The solution scales to your document volume, and smaller repositories process faster and at lower cost. The compliance requirement is the same regardless of size, and the architecture adapts to match.
The solution runs entirely on your internal infrastructure. No document, Aadhaar number, or metadata leaves your server at any point. Masking happens on-premises, and the system has no outbound network dependency.
Conclusion
Planning something similar?
Art Housing Finance came to HabileLabs with defined requirements: 99% accurate Aadhaar masking, secure offline deployment, DMS integration, and full compliance within one month. The engagement delivered all of them. For housing finance companies and BFSI institutions handling daily document volumes under RBI obligations, it shows what is achievable when accuracy, security, integration, and timeline are treated as equal constraints rather than trade-offs.