What can gaming companies learn from artificial intelligence analysis of Gamer 1.5M conversations | Creativ

Creativ Today it appears as a new type of market intelligence company. AI is used to conduct feelings analysis on 1.5 million conversations about the best publishers and titles of the game.

This means that it uses artificial intelligence to find out what players think about 17 of the best publishers of the game – with ideas produced by machine learning. Creativ’s AI analyzed more than 1.5 million online conversations via Reddit, YouTube, Discord and News Ponsts for six months. It took about 10 days to do it. The company has gone through about 9300 news and feeling about gaming publishers. Then she analyzed her study, which covered the period from November 1, 2024, until the end of April 2025.

I talked to Creativ Win Morton CEO, CTO JOE LAI and CIIO (Vibhu Bhan) Vibhu Bhan about results. Here is some information from the exclusive analysis.

“We call ourselves a self -marketing company,” Morton said. “Essentially, the reason we are studying is that we invented how to do consumer visions using LLMS,” said Morton. “The study is literally the swallowing of a million and a half of the consumer talks on these different publishers.”

In extracting “feelings analysis”, the goal is to know what players in gaming companies think based on what these players say on social media. Behehan said that the AI ​​Large Language (LLM) was trained to discover mockery, the game of the game, and more nuances.

Behehan said: “The real innovation here is a better understanding of the context and colloquial horizons, so the analysis of feelings is more context and not just a degree.” “If you look at the analysis of traditional feelings, it looks at the existence of certain words. But the language is complicated.”

Feelings analysis have appeared in recent years as a way to understand the age about a game or company. But the analysis often suffered because the analysis used did not really understand the players or their comments on themes. Now, with LLMS, Morton said that machine learning understands the complex nuances and does better through more data you can absorb.

In one examples, Creativ found that the fans were not happy when actor Henry Cavill was expelled from the leadership role of Giralet in the Witcher TV program on Netflix. Basically, Netflix should not have launched Cavill, as it led to a comprehensive negative impact on the Witcher concession. It turned out that the show affected the total feelings, instead of the video game series.

The company accommodates the data and then comes with dozens of feelings about the publishers of the game to find out what they did to help or harm their brand recently in a conversation with the players. Old reports can know the number of times a set of words (such as a game or company name) has been used. But often he did not have the ability to understand the full context about a discussion about games and then summarize them properly. But LLMS is better in understanding the context about a large amount of data.

Lay said: “The context becomes more important because this allows you to understand the direction of feelings because there may be two topics in the sentence. The second thing is this key that we do because of ridicule, which is seen as a false positive when it is a negative reaction.”

Lay said that LLMS has a better ability to understand the context of the language.

“The beauty of LLMS is that we are able to collect and train our models on these games data,” said Lay. “We are able to train models to be able to discover the news line that appears for each of these games, as well as if they are used in a positive or negative way.”

The largest talks for the conversation

One of the things that LLMS picked up is that players have strong opinions on the exclusive, and whether the platform owner should keep the best exclusive game or transfer this game to other platforms in order to generate more sales. Fans who invested their money in a specific control unit did not like it.

The biggest topics of the conversation included the liquefaction of the game, privileges, gaming platforms, privacy, and the unification of industry and companies. Upon liquefy, the players have rewarded open communications on the rules and studios that avoid income models that affect play and mechanics. This was the broadest direction in the data set, where consumers realize that Activision Blizzard, UBISOFT, EA, Amazon, Netease, Evolution Gaming and Roblox as two bad criminals for poor liquefaction practices.

In addition, LLMS picks up the natural conversations. In contrast, the study puts the player on alert that they are explored for their opinions. The player may think about whether he will answer honestly or not, based on what they believe that the study researcher wants to hear.

How are the companies most likely?

Netflix has not had a lot of history as a gaming publisher, and her mobile games have not yet been huge. This helps explain the reason for obtaining a negative degree of players. Some feelings happen about a game, just like the American Professional League game, but many of them happen outside the game on social media.

Morton said the games get a great awareness of Hollywood, as films that depend on games such as Minecraft and TV show, Last of Us gets high reviews and access to more people who do not know games.

“The great part of this technique is that you can move specifically to what makes people happy and sad,” said Morton.

Activision Blizzard had a lot of gossip on World of Warcraft. But many players were not admired how the company dealt with the move from Overwatch to Overwatch 2. Ubisoft also came out with the worst degree of all the publishers of the game, but it was not clear. He had a lot of discussion about Assassin’s Creed: Shadows. But that game received positive reviews unlike previous games such as Star Wars: Outlaws and Skull & Bones.

For this study, the company did not focus on any specific game. But you can do it in the future.

With LLMS, the study can be conducted in 10 days, compared to weeks for other methods. Morton said that LLMS can absorb data and broadcast it faster, but it can analyze more data and much faster. Over time, the analysis can get a lot of granules, focusing on any specific personalities or other details. Such an analysis can give the team an opportunity to reach other personalities if it has a negative degree.