Competitor Analysis with an AI

Using AI Effectively in Game Business

Working in game business operations often feels like entering a battlefield each day. Whether you’re a game PM, business planner, operator, or marketer, we all share the same goal:

Read the market fast, stay on top of trends, and communicate our game clearly.

However, tasks like “competitor analysis,” “understanding business model (BM) structures,” “report writing,” and “communication with development teams” are essential but hard to delegate. Often, they end up as late-night solo work.

Today, AI has evolved enough to support these real-world tasks.

From analyzing competitor games to detecting real-time market trends, once you understand a bit of structure, you can start using AI tools much more effectively.

ℹ️ Competitor Analysis — Achieving Both Speed and Depth with AI

Let’s say you’re preparing for a global launch of a new MMORPG. A typical approach is to identify successful competitors based on criteria like:

“MMORPGs released globally in the past year with over 1 million downloads.”

You might collect data from Sensor Tower, data.ai, YouTube trend videos, or Steam DB.

But what you really want to know is:

Why did users love this game?

What kind of user experience (UX) and monetization design kept them engaged?

Playing the game, analyzing it, and writing a report takes more time than you have.

This is where AI can truly help.

① Finding a Trustworthy AI Research Assistant

Tools like Perplexity.ai serve as AI-powered search assistants.

With the right prompt, it provides real-time insights: launch dates, features, user reactions, and key metrics, all in a structured summary. It even includes citations for credibility and suggests related questions to help you understand the broader context.

Perplexity displays source references in gray numbers like [6], which you can click for verification

Unlike ChatGPT, which responds based on pre-trained knowledge, Perplexity pulls real-time data from the web — a big difference when accuracy and freshness matter.

However, as with all generative AI, hallucinations can occur, so you should always verify the information.

② Detecting Real-Time Trends — Timing Is Everything

In the game market, timing is a key competitive edge.

Game business teams need to respond quickly to competitor campaigns or updates. While Sensor Tower and data.ai offer solid post-analysis tools, there’s often a time delay.

To solve this, you can use Google Trends and automate monitoring with a simple tool like “pytrends”.

“Pytrends” is a lightweight Python tool that doesn’t require an API key. It automatically pulls real-time keyword trends from Google, search volume, and related terms. Also, non-technical users can easily install and use it with minimal setup.

With pytrends, you can spot spikes like:

“Search volume surged after Game A’s Season Pass release,” or “Negative keywords increased due to Game B’s monetization controversy.”

These insights help you adjust your own BM design or user response strategy.

If you have an in-house data team, you can even integrate pytrends into a more advanced analysis system.

③ How to Analyze Competitor In-Game Content

The best way to understand a competitor is still to play their game.

But when time is limited, what can you do?

That’s where a case shared at a recent conference comes in: HUDstats.

HUDstats originally began as a platform for analyzing real-world sports matches, particularly player movements in soccer. However, during the pandemic, the company rapidly expanded into the eSports space. The core idea behind the technology is simple: with just gameplay video, the AI can automatically perform analysis

Key features include:

Real-time player tracking: location, movement, combat, skill usage

Automatic detection of turning points: kills, objectives, wins

Heatmaps and behavior pattern visualization

It also generates content such as:

Highlight summary videos: Automatically generates highlight reels by collecting the most important moments of the match

Event timeline: Organizes key in-game events into a timeline format for easy reference

Rankings and statistics: Visualizes player movement paths and combat locations, along with various statistical data

Auto-generated commentary and visuals: Creates text and image-based summaries of major moments, ready for use on social media

HUDstats uses generative AI to automate live content creation and real-time commentary

What’s the Connection Between eSports Video Analysis and Competitor Game Analysis?

At first glance, eSports commentary and competitor game analysis may seem unrelated. But they share common goals. When analyzing competitor games, the objectives typically fall into three main categories:

Competitor understanding: Analyze the core gameplay mechanics, UX flow, monetization points (BM triggers), and key UI/UX transition timings

Pattern detection: Automatically identify recurring loops such as quests/missions, ad trigger timings, and reward delivery

User behavior and emotion prediction: Predict when players are likely to drop off or where engagement spikes during gameplay

Viewed through the lens of gameplay video analysis, competitor analysis can also be automated, if you have gameplay footage.

With the right tools, you can identify:

In the context of full video analysis, this means that competitor analysis can be largely automated, if you have gameplay footage. From this footage, AI can identify elements such as retention points, monetization-driven UX flows, or even highlight segments that indicate peak user engagement.

Beyond that, your own live ops or PM teams could use updated videos from your game to detect potential balancing issues early in the gameplay flow, before user complaints even arise.

ℹ️ How Does Video Analysis Work, and Can It Really Generate Content?

Let’s briefly look at a quick technical note, how gameplay video analysis works, and how it can even lead to automated content creation.

First, AVA (Ascendent Video Analytics) is used to extract objective data from the gameplay footage. Then, technologies like LLMs and RAG help transform that raw data into structured content and narrative — playing a key role in automating storytelling.

So what exactly is AVA?

Put simply, it’s an AI system that reads on-screen elements — like kill counts, timers, icons, colors, and player movement — based on numerical and positional data. Think of it as a smart AI cameraman watching the game. Without human input, it can recognize and output: “Team A just got 3 kills.”

From there, it automatically identifies:

Who’s fighting and where

Which skills are most frequently used

Where players tend to gather on the map

Which content segments disrupt the user flow

This means you don’t need to play the game or manually review every second of footage. With just gameplay videos, you can extract key insights such as “Frequently used skills”, “Dominant characters,” or “Points in the game where users cluster or drop off.”

Finally, once the analysis is complete, the AI can auto-generate content. Based on your chosen template and language style, a full report can be created, allowing us to complete our work much more efficiently.

♻️ Final Thoughts – It’s Time to Share the Load with Your AI Assistant

AI is no longer just a “technology.” It has become a practical tool that helps game business professionals save time. Today, AI can support tasks like analysis, content creation, and competitor comparison, all of which used to be manual and time-consuming.

And importantly, we, game business stakeholders, don’t need to be technical experts to use it.

Core game data can already be explored and summarized using open AI tools like Perplexity. Gameplay analysis can be done with just video footage—through tools designed to work without requiring development access.

AI plays the role of a powerful assistant that helps organize complex data and insights.

That leaves business planners, PMs, and marketers free to focus on interpreting the data and making informed decisions.

In the next post, we’ll explore practical examples of how AI can reduce repetitive work during the early phases of game publishing, from market validation to launch execution.

※ 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