PKFARE Partners with Kiro to Redefine AI-Era Software Engineering — Powering Travel Distribution Innovation

PKFARE is a leading global B2B travel marketplace, operating in 100 countries and regions across Asia-Pacific, Europe, North America, the Middle East, and Latin America. By continuously expanding its global network of suppliers and distributors for flights and hotels, PKFARE helps airlines, hotels, and 3^rd party suppliers extend their market reach, while enabling travel sellers such as OTAs and TMCs to efficiently source high-quality travel offerings. PKFARE's core competitive advantage lies in its ability to deliver more competitive and relevant content through a robust supplier network and advanced technological capabilities. Entering a critical growth phase, PKFARE is accelerating expansion through a dual engine of technology enablement and operational efficiency. This strategy requires a scalable technical architecture capable of supporting massive concurrent global transactions while dramatically improving software delivery efficiency.

PKFARE aims to establish an AI-driven, high-efficiency operating model that enables its technology teams to shift from manual execution to AI-native innovation, while cultivating an "AI-first" culture. As its global business continues to expand, the company faces two key technical imperatives: first, building high-performance, stable business systems that support agile deployment for global users while enhancing security, compliance, service quality, and operational efficiency; and second, reducing operational complexity through generative AI, lowering development barriers and allowing testing, operations, and other non-technical personnel to contribute to tool and automation script development. For PKFARE, this represents not only a pathway to support business growth but also a strategic opportunity to multiply human productivity and achieve a step change in core competitiveness through a fundamental transformation of its software engineering paradigm.

"In our view, Kiro is more than just an intelligent tool—it is an efficient, smart, and highly reliable digital partner. Leveraging Kiro's leading spec-driven development paradigm, our test and development team independently delivered a commercial-grade Technical Support Ticketing System in only five weeks, without relying on any traditional R&D resources. The system now exceeds 100,000 lines of code and supports daily operations for nearly 350 employees worldwide, earning strong recognition from users."

Konson Zhao
CTO, PKFARE

To meet the evolving technical demands of global expansion, PKFARE partnered with Amazon and adopted Kiro as a core tool for its R&D team. As early as 2024, PKFARE had proactively leveraged Amazon Q to improve foundational coding efficiency, marking its initial exploration of generative AI in software development. Following the release of the Kiro preview in 2025, PKFARE became one of its earliest adopters, rapidly applying it to more complex development scenarios. In parallel, PKFARE's technology team remained closely aligned with emerging trends in AI-driven software engineering through AI journals regularly published by Amazon. This sustained technical investment and ongoing exposure to frontier AI insights have strengthened PKFARE’s conviction: Kiro has evolved from an AI-enabled IDE into a powerful engine for shifting the team’s mindset from manual execution to AI-native innovation.

To fully leverage Kiro as an Agentic AI IDE for spec-driven development, PKFARE embeds it deeply across the entire product and R&D lifecycle—from requirements definition and system design to coding, unit testing, and application deployment. The team adopts a strategic "coarse-to-fine" approach: for coarse-grained requirements, such as internationalized display features including language-switch buttons, hierarchical field translations, and error message handling, Kiro leverages extensive project references to automatically complete details. For fine-grained requirements involving complex logic, such as refactoring scheduling calendars, product and operations teams prepare well-structured business rules documentation to guide Kiro in generating standardized code. This approach unifies communication standards across product and R&D teams, significantly improving the consistency and compliance of product documentation.

During system construction, Kiro's spec-driven development approach proves critical for PKFARE. By establishing a structured collaboration model consisting of "requirements.md", "design.md", and "tasks.md", it effectively mitigates the limitations of AI in understanding context, enabling AI to generate optimal solutions based on current requirements and historical designs. During the integration of platform entry points, domain-driven design (DDD) principles guide Kiro to consolidate fragmented platform entry points into a unified onboarding workflow while maximizing backend logic reuse.

The PKFARE team gave Kiro a perfect five-star rating for developer experience (DX). By combining local debugging techniques—such as one-click script execution, domain terminology accumulation, and real-time error correction—with Kiro's ability to handle complex task chains, PKFARE team members can quickly ramp up and deliver commercial-grade results.

Completed a 100,000-line Technical Support Ticketing System through development, testing, to go-live within five weeks

With Kiro's spec-driven development approach, PKFARE validated the feasibility of building complex systems from scratch using AI—marking a foundamental upgrade in the R&D paradigm and a step change in engineering capabilities. While conventional single-turn conversational AI struggles to handle complex context, Kiro leverages structured "requirements.md", "design.md", and "tasks.md" to deeply understand historical designs and accurately address new challenges. As a result, a three-member team independently delivered a 100,000-line Technical Support Ticketing System in just five weeks. The team achieved an average output of 33,000 lines of code per person, while keeping the average bug count below 15 per 10,000 lines of code—demonstrating both high-quality delivery and exceptional R&D efficiency.

Reduced test case authoring and execution time from three days to half a day, with over 75% test case usability

Kiro also delivers significant improvements in quality assurance and process standardization. By embedding Kiro across the testing lifecycle, previously labor-intensive test case authoring has become highly automated. Kiro demonstrates strong logical reasoning capabilities across complex scenarios, and its AI-generated test cases boast a usability rate of over 75%. With Kiro’s assistance, the time required for test case authoring and execution dropped from three business days to just half a day — representing nearly a sixfold increase in delivery efficiency and freeing test engineers to focus on higher-value activities such as system quality control and risk identification.

Accelerated innovation projects from requirements alignment to first prototype release by 60×

In terms of innovation speed and agile responsiveness, Kiro serves as a highly competitive "acceleration engine" for PKFARE, fundamentally reshaping the trial-and-error and delivery cadence of projects. For new business models and proof-of-concept (POC) initiatives, teams can rapidly transform ideas into executable minimum viable products (MVPs) with Kiro's support. Previously, projects of a similar scale required three to six months to progress from requirements alignment to first prototype release. With Kiro's deep understanding and automated system construction capabilities, initial prototypes can now be completed in just three days. This acceleration in prototype development boosts innovation and iteration efficiency by 60 times, enabling PKFARE to capture first-mover advantage in the highly competitive global travel market at a lower cost.

Looking ahead, PKFARE will continue to deepen its strategic collaboration with Amazon in generative AI. Plans include tighter integration between existing internal tools and Kiro, as well as connecting Kiro to PKFARE's private business knowledge base to enhance Kiro's understanding of business scenarios. Moreover,  PKFARE is committed to fostering an AI-first R&D culture by refining coding standards, introducing high-quality data sources, and further unlocking Kiro's potential in complex task automation. Through long-term collaboration with Amazon, PKFARE will continue to drive innovation through technology and work with global partners to build a more efficient, transparent, and dynamic travel distribution ecosystem.

Kiro helps you do your best work by bringing structure to AI coding with spec-driven development.