The tidal wave of AI currently crashing through the IT sector shows no signs of receding, threatening to reshape the industry's entire landscape. As companies scramble to figure out how to ride the current, NEXUS Co., Ltd. and KRAFTON have stepped up to share their own messy, real-world experiences.

On the third day of the NEXON Developers Conference 2026 (NDC26), held on the 18th at the Gyeonggi Content Agency's international conference hall, a session titled 'The AX Journey of NEXUS Co., Ltd. and KRAFTON: What We Tried and What We Abandoned' took center stage. Under the moderation of Professor Kim Sang-kyun, both industry giants laid bare the trials and errors of their respective AI Transformation (AX) journeys.

The 'AX central organization' within a game company usually emerges from a collision between grassroots demand and executive mandate. At NEXUS Co., Ltd., this evolved from groups like the Intelligence Labs, which initially handled live service support and automation, into today’s 'AI Division.' KRAFTON, conversely, took the 'AI First' route after a top-down declaration, establishing an 'AI Transform' task force to dismantle technical bottlenecks for the rank-and-file. These hubs now act as the primary control towers, laying the infrastructure for the wider AX transition.

While both companies set up new groups, their strategic trajectories were polar opposites. NEXUS Co., Ltd., with its massive workforce and projects, leaned into a 'bottom-up' philosophy rather than enforcing a rigid standard. Division Head Kang explained, "It might sound a bit irresponsible, but we really focused on the 'let’s just try it' mindset. We poured our efforts into AI literacy education and spread successful use cases across teams. That’s how we identified the 'champions' who led the charge from within their respective departments."

Despite these contrasting approaches, both companies converged on the same objective: boosting employee AI proficiency and uncovering practical use cases. The tangible results are already visible. NEXUS Co., Ltd. reported that by empowering their live service teams to handle everything from data extraction to console construction via AI support, they have drastically reduced response times for live issues.

KRAFTON highlighted how even non-developers, such as HR staff with zero coding experience, leveraged AI tools to build automated recruitment management systems. Both leaders agreed that building a track record of small, successful wins is the critical first step in spreading AX adoption.

However, success is always flanked by scars. NEXUS Co., Ltd. candidly discussed the failed attempt to implement 'OpenClo' company-wide. While it seemed like a powerful tool for local machine control, the blistering pace of updates, persistent security headaches, and spiraling infrastructure costs forced them to abandon the project. KRAFTON shared similar growing pains, noting that adopting massive, enterprise-grade AI tools often backfired when those tools failed to account for the unique, granular requirements of game development.

What were the common hurdles they faced after these failures? Surprisingly, both agreed on the same culprits: employee fatigue and the management of 'AI usage costs'—specifically, token fees. The relentless influx of new technology has left staff exhausted, hindering both efficiency and adaptation. NEXUS Co., Ltd. is currently preparing a 'long-term AX roadmap' to help predict and mitigate these fluctuations.

The financial reality is impossible to ignore. Both companies are now working to optimize costs. They are researching 'cost prediction models' to ensure every AI token spent translates into genuine productivity gains. KRAFTON has gone a step further, launching a cost dashboard and a price-performance index for LLMs, allowing internal teams to choose the most efficient model for their needs. Both firms are also developing proprietary foundation models to hedge against the risk of future monopolistic price hikes.

The discussion concluded with a deep dive into how game companies can secure future competitiveness in the AX era. Both leaders defined the gaming industry as a 'blessed domain'—one far more agile and open to AI experimentation than heavily regulated sectors like finance or commerce. By delegating repetitive chores like asset generation and source code validation to AI, studios can slash overhead and iterate prototypes with newfound speed.

Ultimately, competitiveness in the AX era boils down to the power of focusing on what only humans can do. When AI handles the grunt work, developers must double down on the essential questions: 'Why is this game fun?' By viewing AI not as a replacement, but as an 'augmentation tool' that expands creative horizons, developers can shed their fear and evolve into true AI-native creators for the next generation of gaming. Adieu.