Real User Insight with an AI

Focusing on Insights, Not Repetition

Anyone involved in game development or live service operations knows how easy it is to lose sight of what truly matters when things get busy. Situations like these may sound familiar:

“Let’s skip the test code for now and just get the feature working…”

“There’s a flood of comments on the trailer. Who has time to read them all?”

Now imagine if those repetitive, time-consuming tasks could be handled for you. It would free up time and mental space to focus on strategic decisions and more creative work.

This is exactly where EA (Electronic Arts) is focusing its efforts — empowering developers, QA teams, live ops teams, and game business professionals by letting AI take over the “repetitive but essential” tasks first.

ℹ️ Test Code, Written by AI

In game development, “unit tests” are essential for automatically verifying that core features work as intended. Yet under tight schedules, writing test code is often deprioritized.

To address this, EA integrated AI-powered test code generation into its internal tools. After completing a feature, developers can simply click a button, and the AI generates a draft unit test on the spot.

For example, if a developer builds a feature that converts emotions into emojis, the AI creates a test that validates whether the function returns the correct emoji for each emotion.

This not only saves developers time but also helps catch basic issues before QA even begins—allowing them to focus more on building complex and creative features.

ℹ️ Hundreds of Comments? Let AI Summarize

When a game trailer is released or a beta goes live, marketing and live operations teams need to quickly assess community response. But manually reviewing and organizing hundreds of comments from YouTube, Twitch, or forums is both time-consuming and inefficient.

EA has started using AI to streamline this process.

By inputting a social media link into an internal tool, the AI reads the first 100 comments and summarizes both positive feedback and suggested improvements. For example:

Positive: “The F1 game graphics and sound are amazing.”

Suggestion: “The AI opponents are too easy, so there’s no real challenge.”

Rather than just labeling feedback as good or bad, the AI identifies what users liked and what needs improvement—turning raw reactions into actionable insights.

3️⃣ AI Supports User Testing Too

AI is also making an impact in user testing, including Focus Group Testing (FGT) and Focus Group Interviews (FGI).

In one Nexon project, AI analyzed real-time video of players during gameplay to track facial expressions and detect moments of immersion, confusion, or enjoyment. This enabled both developers and business teams to quantify emotional reactions that traditional surveys often miss.

(Source: FGT Facial Analysis Framework, Kwon Seung-jin, NDC 2021)

ℹ️ Streaming Analysis: Tracking Eyes and Reactions

On platforms like Twitch, eye-tracking data helps identify which on-screen moments capture viewer attention, and which tags or entry points lead to engagement.

AI can also map viewer acquisition paths—for example, whether a user arrived via a Twitter link, homepage recommendation, or clicked on a “new game” tag. This journey is automatically captured and organized to reveal engagement drivers.

In addition, gaze analysis can identify areas of focus in UI or video content. Eye-tracking technology tracks the viewer's eye movements to measure how long they stay at specific locations on the screen.

For instance, if a viewer's gaze lingers for only 2 seconds out of 30 on a UI displaying character stats, that element likely didn't receive enough attention. These findings are visualized in heatmap format, allowing quick identification of focal points and neglected areas.

In addition, gaze analysis shows which parts of a UI or video received meaningful attention. Eye-tracking tools measure how long viewers focus on different screen elements. For instance, if the viewer glances at a character stats panel for only two seconds out of thirty, that element likely failed to attract interest. Heatmap visualizations make these patterns easy to interpret, highlighting what stood out and what was overlooked.

Infrared heat map changes according to the gaze of Twitch viewers

This offers more than click-based analytics — it reveals “real-time emotional engagement,” providing deeper insight for streaming content and broadcast-style game design.

ℹ️ Automating Focus Group Surveys

AI is also transforming how post-play surveys are conducted in Focus Group Interview (FGI).

Traditionally, when players join a Focus Group Test (FGT) or FGI, game teams distribute surveys manually and process the results by hand. Now, AI chatbots can automatically send surveys, and participants’ free-text responses are analyzed using natural language processing (NLP) techniques. For example:

“The tutorial is too long.” → Tagged as negative sentiment under “UI Onboarding”

“Combat was fun, but the rewards felt weak.” → Positive for “Combat,” improvement needed in “Reward System”

By tagging and quantifying open-ended responses, QA and operations teams can quickly identify key feedback trends without having to read through lengthy comments.

Repetitive analytical tasks are reduced, freeing game business stakeholders to focus on drawing real insights.

♻️ Final Thoughts

Across all these examples, the message is simple:

AI is not amazing because it’s powerful. It’s helpful because it takes care of repetitive work that distracts us from what really matters.

By offloading routine tasks —like test code generation, comment summarization, and feedback tagging — AI allows developers, marketers, and UX teams to focus on higher-level decisions and creative problem-solving.

Rather than replacing people, AI supports them by freeing up time and mental energy for meaningful work.

Whether it’s generating test code for developers, summarizing user comments for marketers, or tracking emotional reactions for UX testers, AI is becoming part of our workflow.

In the next post, we’ll explore how AI is supporting the game localization process, a key area in expanding into global markets.

※ Disclaimer: This content reflects the author’s personal views and includes only publicly available examples. It does not represent the official position of any company mentioned