In this case study, LTS Global Digital Services (GDS) leveraged Valitract’s AI-powered document extraction platform to automate complex multilingual document intake, accelerate annotation workflows, and improve processing accuracy for high-volume enterprise projects worldwide.
18%Accuracy Improvement
64%Faster Processing
66%Cost Saving
3 daysAnnotator Onboarding
Client Overview
Company
LTS Global Digital Services (GDS)
Industry
AI Data Annotation, Data Entry & RPA / BPO
Headquarters
Hanoi, Vietnam
Founded
2015
Scale
500+ annotators, 200+ projects delivered
Geographic Reach
11 countries across Asia-Pacific, North America & Europe
Interview Subject
Mr. Giang, Chief Operating Officer
LTS Global Digital Services (GDS) is a leader in delivering large-scale data annotation, data entry, and Robotic Process Automation (RPA) services to enterprise clients across 11 countries.With a workforce of more than 500 skilled annotators and a portfolio of over 200 completed projects, GDS has built its reputation on a quality-first philosophy backed by ISO 27001 certification. The company serves demanding industries including automotive, BFSI, manufacturing, construction, and retail.
The Challenge
As GDS expanded its annotation and BPO operations for a growing global clientele, the company faced increasing challenges with document intake. This process is the essential first step before any annotation or data entry can begin. Client projects arrive with a wide variety of source documents: instruction briefs, labeling guidelines, classification schemas, delivery forms, vendor invoices, and data manifests.At the same time, enterprise clients were increasingly demanding high-volume projects, sometimes requiring hundreds of thousands of labeled items delivered within just 48–72 hours.The pressure of high-volume, time-sensitive projects transformed these operational friction points into four critical challenges:
Low Extraction Accuracy: GDS’s existing OCR tool achieved only 79% accuracy when processing mixed-format client documents, including handwritten project briefs, scanned instruction PDFs, and multilingual annotation guidelines. This resulted in QA teams having to review and correct nearly one in four documents manually.
Slow Project Kick-off: On average, 22 hours elapsed between a client submitting a project brief and annotators receiving a correctly parsed task sheet, creating frustrating delays, especially for time-sensitive LLM training batches.
High Rework Costs: A 28% QA rework rate meant dedicated staff were spending the majority of their time correcting extraction errors rather than delivering value, with a per-document processing cost of $2.80.
Annotator Onboarding Friction & Scalability Constraints: Scaling the workforce was hindered by an unintuitive interface that required 4–5 weeks of training for new annotators. This prolonged onboarding cycle meant GDS could not rapidly increase headcount in response to project surges, leading to higher recruitment costs and missed growth opportunities.
“We pride ourselves on quality. But the irony was that our own document intake process was one of the biggest sources of errors in every project. We needed a platform smart enough to handle the diversity of documents our clients submit, ranging from clean digital PDFs to handwritten notes scanned on a smartphone.”— Mr. Giang, COO, LTS Global Digital Services
The Solution: Valitract AI Data Extraction Platform
Following a structured evaluation of four competing platforms, including tools from well-established OCR vendors, GDS selected Valitract for three decisive reasons:
Purpose-Built AI Accuracy for Diverse Document Types
Valitract’s machine learning models are trained on an exceptionally wide range of document formats, languages, and quality levels, making it uniquely suited to GDS’s heterogeneous client document landscape. Out-of-the-box accuracy reached 97% across project briefs, labeling schemas, classification templates, delivery confirmations, and vendor invoices, including documents in Vietnamese, Japanese, Korean, and English. The platform’s contextual understanding correctly handled handwritten annotations and low-resolution scans that had previously defeated GDS’s legacy tool.
Intuitive Interface Designed for Non-Technical Operations Teams
Although GDS’s annotation workforce is operationally skilled, it is not trained in software engineering. Valitract’s clean, visual dashboard, featuring document previews with highlighted extraction fields, one-click corrections, and drag-and-drop batch uploads, was designed with this user profile in mind. Key capabilities that resonated with the GDS team included:
Side-by-side document view with color-coded extracted fields for fast verification
Smart auto-suggestions that learn from annotator corrections over time
Bulk task assignment linked directly to project management workflows via API
Real-time processing dashboards with per-client accuracy and throughput metrics
These features allowed GDS to reduce onboarding time from weeks to just three days, empowering new staff to contribute productively and independently almost immediately.
Seamless RPA Integration and Transparent Pricing
Valitract’s open API integrated cleanly into GDS’s existing RPA pipelines, enabling end-to-end automated document routing: from client submission, through extraction and validation, to task assignment in GDS’s annotation management system, with no manual hand-offs. The per-document pricing model provided full cost visibility and eliminated the unpredictable infrastructure expenses of the legacy on-premise solution.
Implementation Process
The deployment was completed in four weeks with active support from Valitract’s solutions engineering team:
Week 1: Integration & Piloting: API connection established with GDS’s RPA workflows and annotation management system; 200 representative client documents across 6 languages used for benchmark testing.
Week 3: Staff Training & Parallel Running: All team leads completed 3-hour training sessions; Valitract and legacy system ran in parallel for cross-validation.
Week 4: Full Production Cutover: Complete transition to Valitract with Valitract support team monitoring performance metrics and providing same-day issue resolution.
Results & Business Impact
Six months after full deployment, GDS has seen transformative improvements across every key operational metric:
Metric
Before Valitract
After Valitract
Document Processing Accuracy
79%
97%
Avg. Turnaround per Task Batch
22 hrs
8 hrs
Cost per Document Processed
$2.80
$0.95
Annotator Onboarding Time
4–5 weeks
3 days
QA Rework Rate
28%
6%
Key Business Outcomes
18-point accuracy gain: From 79% to 97%, reducing manual QA rework by 79% and freeing senior annotators to focus on high-complexity tasks rather than error correction.
64% faster project kick-off: Average processing time dropped from 22 hours to 8 hours, enabling same-day task assignment for 88% of incoming client projects.
66% cost reduction per document: Processing cost fell from $2.80 to $0.95, saving $1.85 per document. Based on around 40,000 documents processed per month, this translates to an estimated $74,000 in monthly savings, approximately $890,000 annually.
Onboarding acceleration: New annotator readiness cut from 4–5 weeks to 3 days, allowing GDS to scale headcount rapidly in response to large project surges without quality risk.
RPA pipeline reliability: Zero extraction-caused RPA failures recorded since go-live, compared to an average of 14 workflow interruptions per month under the legacy system.
Competitive differentiation: GDS has achieved a much lower QA rework rate while maintaining a guaranteed 8-hour document-to-task SLA. This combination of precision and efficiency has directly led to winning three new enterprise contracts.
“Valitract did more than just fix a pain point. It changed what we can promise our clients. We now compete on speed and precision in a way we simply couldn’t before. It has become a core part of how we deliver quality at scale, and an important reason clients choose GDS over the competition.”— Mr. Giang, COO, LTS Global Digital Services
Conclusion
LTS Global Digital Services’ experience demonstrates the compounding value that AI-powered data extraction delivers when deployed at the front end of large-scale annotation and BPO operations. By replacing an error-prone, slow, and expensive legacy OCR tool with Valitract, GDS achieved dramatic improvements in accuracy, speed, and cost, while simultaneously eliminating the training burden that had constrained its ability to scale.For companies in the data services industry where accuracy is the product and speed is the differentiator, Valitract’s combination of advanced AI accuracy, intuitive design, and deep integration capability makes a compelling and measurable business case.Ready to transform your document processing operations?Contact Valitract to start your free pilot today. Valitract – Next-gen AI-Powered Data Extraction Platform
