Need a Gaming Coach? There's an AI Agent for That

Ever wanted to be able to summon a gaming buddy on demand? AI can help fill the void—and maybe up your game, too.

Balance.fun has launched as an extension of the E-Pal platform, becoming the home for AI agent gaming companions that will coach and support you while you play your favorite video games. This comes ahead of an Ethereum token launch for the project, which is currently scheduled for late Q1 2025.

E-Pal allows gamers to pay human companions to play video games with them, offer coaching or just someone to chat with while gaming. This platform was created to address the loneliness that some gamers experience as they struggle to sync up with their friends in-game, or simply don’t have anyone to play with.

Balance.fun removes this human element by offering a chat box with an AI agent that will support and coach you through your games. This product is less focused on the loneliness aspect of the E-Pal platform, and more on offering a coaching alternative.

“They can do it in a way that's more efficient, without that kind of emotional aspect of a human—because, you know, sometimes people just don’t get along,” pseudonymous Balance.fun Head of Ecosystem, Tri, told Decrypt, “If you're just specifically trying to improve your game, then I think that's where AI E-Pals can really become useful.”

When creating an AI agent on Balance.fun, the user can customize their personality by assigning it hobbies, jobs, and other traits, as well as select the character’s appearance. These agents are partly inspired and trained on data collected from the “best human E-Pals,” in regards to their interaction styles, service offerings, and gameplay expertise.

The agents use reinforcement learning, which means they learn based on the reward or punishment that comes after an action. So if your companion is watching you play League of Legends, then they may notice patterns in your gameplay that regularly lead to negative results, and can notify you about weaknesses in your game.

In theory, this means that AI agent companions can quickly become experts on any game. That’s an advantage over human E-Pals, who remain more effective at offering emotional support due to that human touch.

“At the moment, I believe that [the AI agent] capabilities are kind of just: You can chat with them, you can get them to do some tasks, and they can learn from these tasks,” Tri told Decrypt. “Future capabilities are for them to be a sort of scalable business, where our gamers and people that are using the Balance platform can register these E-Pals and gain points in our ecosystem.”

Next steps for Balance.fun include an AI agent marketplace that rewards the creators of the companions with points. These points can also be gained through other methods, such as via a League of Legends tournament that will pay out points based on the player’s performance.

'New Levels of Immersion' Coming to Ronin Gaming Network With AI Agent Tool

Developers on the Ethereum layer-2 network Ronin are set to get an AI agent tool to enhance their games later this year.  The integration with NRN Agents will kick off with a “proof-of-concept demonstration” in February before rolling out across the gaming ecosystem. This could bring “entirely new levels of immersion” to games, the creators of the AI tool say. NRN Agents, a research and development lab for AI agents, was created by ArenaX, the developers of fighting game AI Arena. As such, AI Ar...

Once the EPT token launches on Ethereum, scheduled for late Q1 2025, the aforementioned points will be exchanged for tokens. Tri said that the token is intended to “power everything” in the Balance.fun ecosystem, including governance, staking, rewards, and payments.

On top of this, human E-Pals will later have the opportunity to launch their own “fan tokens” via the E-Pal Fans protocol. This platform will be an amalgamation of launchpad Pump.fun—with tokens first being created on a bonding curve—and social app Friend.tech, due to its social elements.

“Existing E-Pals can leverage their communities from the Web2 side, and transition them onto Web3 using this Fans protocol,” Tri finished.

Edited by Andrew Hayward