Papua New Guinea Immigration Authority Cuts Visa Processing to under 4 Minutes Using Generative AI

The Papua New Guinea Immigration and Citizenship Services Authority manages immigration, citizenship, and border security across the country. It processes citizenship applications, ensures compliance, and handles a high volume of visa requests—often receiving up to 1,000 short–term visa applications daily.

ICSA’s previous digital systems lacked validation checks during data entry, resulting in frequent errors, incomplete submissions, and significant delays. Processing times stretched from days to weeks, as officers manually reviewed and corrected each application—leading to backlogs and affecting the country’s tourism and business goals. “We wanted to stimulate economic growth by making it easier for tourists and business investors to come here,” says Wellington Navasivu, chief migration officer at ICSA. “To make that happen, we needed new technology to help us reduce short–term visa processing times by eliminating our manual processes.”

The organization also faced challenges in extracting accurate information from documents due to disparate legacy systems and the variety of formats and standards in visa applications. Inconsistent review processes and limited operating hours further delayed approvals and created gaps in policy enforcement.

To address these challenges, ICSA needed a modern solution that could streamline visa processing, improve data accuracy, and enhance national security. It looked for a system powered by generative AI, combining natural language processing (NLP) and computer vision (CV) to analyze text and images from complex application documents at scale. This initiative represents the first known deployment of generative AI in a government production environment in the Pacific Islands—a breakthrough for the region. Built and trained locally to meet Papua New Guinea’s specific needs, the system highlights the power of homegrown innovation to deliver real-world impact and advance digital transformation across the Pacific.

To modernize its visa processing capabilities, ICSA partnered with NiuPay, an AWS Partner based in Papua New Guinea that specializes in cloud-native solutions for the public sector. Building on an existing relationship—having previously worked together on projects like an online payment gateway, ICSA chose to collaborate with NiuPay again to design a solution that would accelerate processing, improve accuracy, and enhance national security. “We trust NiuPay, and we were very comfortable working with them again to strengthen our visa processing system,” says Navasivu.

Together, the teams developed MAE, a generative AI–powered document processing system built on AWS and integrated directly into ICSA’s eVisa platform. Designed for flexibility, the solution can handle any document type, with configurable rules that non-technical staff can update as policies evolve. “We created a multipurpose, generative AI–driven system for extraction, interpretation, and decision–making on a variety of file and document types,” says James Inglis, CEO and cofounder of NiuPay.

To achieve this, MAE uses Amazon Bedrock for NLP to analyze text, Amazon Textract to extract structured data from scanned documents like passports and letters, and Amazon Rekognition for CV–based photo validation. The system also includes a configurable rules engine that immigration officers can manage without technical knowledge. “We wanted to remove complexity and minimize changes to ICSA’s existing application environment,” says Jason Kurdzinski, chief technology officer at NiuPay.

A proof of concept was delivered within weeks, and the full system went live just 10 months later following successful testing with 100,000 sample cases. At the core of MAE is a central orchestrator—a decision engine that reviews each document, from photos to supporting letters, against predefined rules. “If the engine is 100 percent sure, it grants the visa automatically,” adds Kurdzinski. “If there’s any uncertainty, the application is escalated for human review.”

Today, around 95 percent of short-term visa applications are processed automatically by MAE, which runs 24/7 and supports continuous operations beyond office hours. Applications that once took weeks are now approved in as little as four minutes, provided all documents are in order. “Our average short-term visa processing time is down to four minutes,” says Navasivu. “We’ve received very positive feedback from some of our largest investment customers and other enterprises with employees needing to travel urgently on business.”

By automating document analysis and approvals, ICSA has increased visa processing efficiency by 50–99 percent and can now handle up to 1,000 applications a day—most without human intervention. This allows immigration officers to focus on complex cases or applications requiring additional scrutiny.

"Working with AWS and NiuPay, we are reducing fraudulent document submissions with MyICSA Analysis Engine, which helps us quickly analyze applications and provide accurate reports to immigration officers," says Navasivu. "By using generative AI on AWS, we’re strengthening our border security by improving compliance checks and significantly reducing the risk of immigration fraud."

Only accurate, up-to-date documents and photographs are accepted, ensuring higher data quality and seamless ingestion into border management systems. Strict data protocols also help keep applicant information secure and compliant with national and international regulations.

Within the first three months, the system processed over 14,000 applications, comprising more than 54,000 supporting documents. Encouraged by the strong results, ICSA plans to extend MAE’s capabilities to include work permits and passport processing.

“As far as we know, this represents the first deployment of AI in a government production environment within the Pacific Island region—a milestone we are incredibly proud of,” said John Rosso, DPS, MP, deputy prime minister and minister for Lands, Physical Planning & Urbanization, Immigration and Border Security. “These advancements position Papua New Guinea as a regional leader in government digital transformation, making our immigration processes quicker, more transparent, and secure.”

Looking ahead, NiuPay plans to work with ICSA to extend MAE to long-term visa applications. The company is also exploring opportunities to bring the solution to other markets. “We designed MAE to solve immigration challenges,” says Kurdzinski. “And we’ve seen strong interest from other potential customers looking to adopt it.”