GGWP, which has used AI for moderating player behavior in games, has expanded its offering with GGWP Pulse, a sentiment analysis tool which is another cool use for AI.
So far, GGWP has used its AI tech for making sure players behave well in online games and to build healthy communities for brands. But now GGWP is a new product which delivers a real-time insights and sentiment analysis, said Dennis Fong, CEO of GGWP, in an interview with GamesBeat.
Similar tools measure the temperature of a brand or game on social media. These “social listening” tools sometimes couldn’t absorb much data. Or they might focus only on a platform like Reddit, which tends to have hyperbolic responses that are not necessarily representative of the vast majority of users, Fong said.
But with the improvements in natural language understanding of AI in recent years and the volume of data that can be processed now, the sentiment can be measured much more accurately in the age of AI, Fong said.
“Social listening tools would crawl Twitter and Reddit and then try to glean insights from those posts that people are making,” said Fong. “A lot of publishers still use those today. Or they just go directly to Reddit to suss out how the community is feeling. The challenge is that it’s a manual process.”
He added, “Or, perhaps more importantly, a lot of the comments that you see on Reddit and Twitter are really hyperbolic in nature, and they also tend to skew really negative, right? If you just listen exclusively to the user’s complaints from Reddit, you may be listening to only 1% of your actual game player community. They’re the loudest, and they’re usually the most hardcore. They may lead a lot of studios down the wrong path about what to iterate on.”
He said the tool provides companies with unprecedented business intelligence they can act on. As a part of the company’s mission to become an end-to-end Community Copilot, Pulse turns previously unseen and unheard in-game conversations into actionable insights, community growth, and revenue generation, he said.
GGWP analyzes anonymized text and voice chat data in real-time to give developers more accurate sentiment analysis and deeper insights into what their players truly think about their games. These insights lead directly to product iteration and can drive real, measurable ROI.
“Our new product unlocks the next level of serving communities,” said Fong. “Specifically, with our GGWP Pulse sentiment analysis, we’re giving companies real time insights into player thoughts and feelings, enabling them to make informed, precise, player-centric decisions.”
“Ultimately, GGWP is moving toward becoming your complete Community Copilot, which will support companies through moderation, automation, sentiment analysis,and revenue optimization,” Fong added.
Several of GGWP’s partners, including thatgamecompany, have already deployed the GGWP Pulse tool, and have been able to immediately identify and address game design elements to provide a better player experience.
“The data we saw from GGWP’s sentiment analysis tool allowed us to zoom in on the map and identify friction inside the game – and address that right away,” said Jenova Chen, founder and CEO of Thatgamecompany, in a statement. “Broad sentiment analysis tends to not allow for surgical improvements; this new offering, alongside GGWP’s safety and moderation tools, takes community and live game management to the next level.”
GGWP Pulse, a next generation business intelligence tool, allows partner companies to:
Access Unfiltered, Real-Time Player Sentiment: Capture the raw opinions and feelings of players while they’re engaged in the game, not weeks after the fact through reviews or surveys.
Cut Through the Noise: Move beyond polarizing, often negative social media posts and see the bigger picture of player experiences and desires.
Spot Emerging Trends: Identify player sentiments and trending topics in real-time, allowing for quicker response times and better-informed updates or patches.
Improve Player Engagement: Use player feedback to enhance the game experience, from fine-tuning mechanics to adjusting community interactions, fostering stronger relationships between brands and players.
On the trust and safety side of monitoring, GGWP has already had to figure out whether someone is speaking in a sarcastic way or is joking. They might drop an F bomb and mean it in an endearing way, rather than in an insulting way. AI is getting better at understanding this context behind speech, Fong said.
“We’re seeing literally evolution — a complete evolution — in a matter of six months,” he said. “The improvements are significant. On top of that, we’ve trained our models based on how people are talking in a game, and we’ve honed it over four-plus years that we’ve been doing business. I think we definitely do that better than anybody else, at least in games. But the sentiment analysis that we’re performing is actually based on the in-game conversations.”
While there is a lot of data to sift through, GGWP already does a lot of it through its monitoring of player behavior. And so from a cost perspective, it can efficiently add the sentiment analysis to the AI processing that is already being done. And by understanding the sentiment of players as they play your game, developers can get honest and unvarnished feedback about what they think of a game or a map or a mode. The sentiment analysis can capture what Fong called “player mood.”
In fact, developers can capture sentiment analysis on a spatial basis. That is, it can understand how players feel when they are in a particular location on a map in a game. One studio discovered a bug in a game because it had negative sentiment from players who went to a particular spot and found that the game didn’t function right at that spot.
Fong noted that his cofounder, George Ng, is an AI expert and he is on the board of the AI lab at UC Berkeley.
“We have real AI experts on our team,” Fong said. “That made our job easier moving into this space.
The company has about 40 people, and a handful worked on the sentiment analysis tool.
A number of developers are using the tech ahead of the launch. We’ll see how the results work out over time.
