Upon ChatGPT’s debut in November 2022, the digital world forever changed. With AI tools in the hands of the public, our lives online would never be the same again. Today, these AI tools have combined with Web3 technology to start a new era of the internet. That innovation is AI agents.
Merging modern AI with the latest blockchain tech, AI agents have made a massive impact on social media, gaming, digital identity and more – and we’ve only just begun.
The impact of AI agents is comparable to that of AI itself – but why are AI agents so important, what are the most notable AI agent projects, and what could AI agents do for us in future?
Here’s our complete 2025 guide to AI agents.
Key Insights
- Software tools that replicate human actions
- Typically powered by, and reward, a token when used
- Made popular in October 2024 by Virtuals Protocol
- Saw major growth in early 2025
- Expected to play a major role in the future of gaming, Web3 and more
What are AI agents?
AI agents are software tools, powered by AI, that train on and replicate human actions and behaviours.
There’s a huge list of ways in which AI agents can be used. Some of the most popular include:
- Social media
- Video games
- Character bots
Social media AI agents impersonate humans on social networks, writing posts, replying to others, and having back and forth discussions. Popular social media AI agents include @jaihoz_, @luna_virtuals, and @aixbt_agent. Many of these agents are ran by, or replicate the behaviour of, infamous names in the community – such as ZachXBT – and are created on online platforms such as Virtuals.
Gaming AI agents can perform human-like video game inputs. This allows users to complete quests, fight enemies and earn rewards whilst away from their devices. Video games with AI agents typically factor them into the ecosystem for that game in a way that benefits both players and developers. Popular gaming AI agent platforms include GOAT Gaming and KGeN, with Ronin set to introduce an AI agent SDK later this year.
Character bots are another type of AI agent. These AI agents adopt the behaviours of a popular character from a game, project or franchise. Users are typically tasked with speaking to these AI agents in an attempt to convince them to perform a particular action, or reveal certain information. A growing number of top projects are experimenting with these types of AI agents, including games such as Blocklords.
To use an AI agent, players pay a fee in the token of that AI agent. AI agents often generate rewards in this token too, motivating their use, experimentation, and optimisation for bigger rewards.
How do AI agents work?
AI agents start are created using popular AI models, such as ChatGPT.
Developers take that AI model, and train it to perform a certain task. Training data could include social media feeds, player inputs for a video game, the lines of a story character, or others. Once that AI model has trained on a sufficient amount of data, it can be released to the public as an AI agent. This usually involves minting a token via an AI agent platform, which users then spend to use that AI agent.
Some platforms require one-time fees to use an AI agent, whilst others operate as a rental service – charging a token fee for a specific duration of time. New services are debuting that host the entire AI agent creation, hiring and deployment flow on a single platform. Popular AI agent websites include Virtuals and Holoworld.
What have AI agents accomplished so far?
AI agents first released to the public in October 2024, when Virtuals launched their AI agent co-ownership service on Base. Virtuals lets users purchase tokens of existing AI agents, allowing them to speculate and profit on the performance of popular AI agents.
Many regard Virtuals as the top AI agent platform, particular for social media AI agents. The site powers @Toshi_base, @aixbt_agent, @RektguyAI and many others. As of February 27, 2025, $VIRTUAL, the token of the Virtuals platform, has a market cap over $700 million USD. AI agent tokens available via the platform have a combined market cap of over $1 billion USD.
AI agent adoption took off in January 2025. In Web3 gaming, Super Champs, KGeN, Illuvium, Nifty Island, Avarik Saga and Ronin all debuted or revealed AI agent products or services. January also saw Avalanche partner with Aethir on a $100 million USD ecosystem fund, with AI agents as the focus. Newer platforms also debuted to expand the AI agent ecosystem, with Holoworld being one of the most notable.
In February 2025, GOAT Gaming debuted their AlphaGOATs AI agents. These are designed for use in their Web3 gaming ecosystem, and was one of the first examples of gaming AI agents being available for public use.
What does the future hold for AI agents?
AI agents are one of the biggest recent innovations of both Web3 and our digital lives.
Whilst some types of AI agents could be considered gimmicks, others have the potential to change the internet landscape as we know it.
So far, gaming is the leading application for AI agents. Gaming AI agents could eliminate the problem of botting in video games, making it a thing of the past. By having AI agents into games, players gain equal access to powerful tools, developers can factor for their inclusion, and all players in the ecosystem are fairly rewarded. This could result in win-win scenarios that rewards both player time and developer ingenuity – rather than the current status quo, where the benefits only go to a few unscrupulous players.
AI agents have made a big impact – and as the concept is still so new, their future looks bright. However AI agents develop in future, it’s a given that they’ll be one of the most exciting aspects of AI, Web3 and our digital lives not just across 2025, but far into the future.
Bullish on web3. Competitive soul.