AI Agents in the Web3 World | The AI Revolution

What are AI Agents?

An AI agent is a digital platform that can interact with its surroundings, gather information, and use that information to perform tasks on its own to reach specific goals. While humans set the goals, the AI agent decides for itself the best actions to take to achieve them.

A simple example of an AI agent is a chatbot. Chatbots can prompt users to gather any information it needs, and use that information to gather relevant data and formulate a solution for a user.

However, AI agents can do a lot more, from powering self-driving cars, to automated trading, and supply chain management. The list is perpetually growing, but what has web3 got to do with any of this?

The Rise of AI Agents

AI agents also carry out tasks, make decisions, and interact with decentralized apps (dApps) and smart contracts on their own, without needing human input. These agents work in a trustless environment, where their actions are controlled by smart contracts and blockchain records, ensuring security and transparency.

Currently, AI agents are able to perform specific tasks independently, but moving forward, this will evolve into something much greater. For example, what Virtuals and their native G.A.M.E. protocol purport to do is create interaction between AI agents, to a limitless degree, which would in essence create swarms of automation that will perpetually compound and create vast value for businesses and individuals alike. (More on this below)

Take a look at this example of AI agents in action! They’re able to communicate between each other to solve problems, and can actually trigger transactions on-chain without the need for human interaction.

Who was the First Crypto Agent?

Truth terminal is considered to be the first AI agent, which popped up on X in early 2024. The agent gained momentum from posting satirical content, with its particular tweet exchange with Marc Andreessen, a venture capitalist, giving the AI $50,000 in Bitcoin. This was the first person-to-AI payment which is what kicked off the launch of the memecoin, Goatseus Maximus “$GOAT”, inspired by Truth Terminal.

Goatseus Maximus reached a market cap of over $1bn. It gained enough traction for Binance to create a listing for it on the perpetual markets. Funnily enough, you can’t buy and hold the token, but you can trade it with 75x leverage.

Use Cases of Crypto AI Agents

Crypto AI agents have multiple use cases, here are a few different ways in which they’re being utilised across web3: 

• Defi Agents – Can monitor and manage various de-fi activities such as yield farming, lending strategies, collateral management and other complex financial strategies. 
• Trading Agents – not only automate trading but can also execute trading strategies in real-time
• Social and Memes Agents – specialise in engaging on social platforms such as X and Tiktok 
• Gaming Agents – offering assistance via personalised game strategies, recommendations and real-time advice whilst gaming
• Content Creation Agents –  agents that create NFTs, music, art and other forms of content.
• Financial Agents – autonomous financial managers for DeFi optimization, including liquidity management, staking, and optimizing yield farming strategies.

Although there is already a wide range of functionality behind AI agents, their uses are growing exponentially, with incredible new features being developed all the time.

The Virtuals Protocol

When it comes to building an AI agent, you may have seen the Virtuals Protocol. Virtuals Protocol is built on the Base and Solana network (More chains to be added soon), allowing users to create, own and deploy AI agents and an attached token. The AI agents built on the Virtual Protocol can interact through test, speech and 3d animations. They are also able to act and perform tasks in games, engage with users on social media platforms or even interact with other AI agents. The Virtual agents can do this because of its G.A.M.E framework.

In many ways, Virtuals laid the foundation for the AI agents wave. It allowed users to create, edit, and trade agents, which led to certain agents being worth upwards of $250 million. But it also enabled cross-platform integration thanks to the G.A.M.E. framework, allowing agents to operate across multiple platforms, which is imperative to global adoption.

G.A.M.E. Framework

On the Virtuals Protocol, the G.A.M.E. framework is described as the decision-making engine used to power agents in different environments and platforms. G.A.M.E’s framework facilitates developers in creating and deploying AI agents, and multiple platforms such as social media platforms, games, and trading environments. 

There are multiple features that make the G.A.M.E. framework attractive. The first is its low code and user-friendly programming kit to help build autonomous AI agents. The next feature is that it’s open for everyone, not only can it be used to build AI agents, but it can also be used on platforms and apps that want to implement AI capabilities. Finally, it’s how customisable it is. Developers can decide what agents can see, what it can do and say as well as its goals. The customisable features and the ability to add various capabilities such as social engagement, on-chain transactions or image generation to name a few.

Agent Commerce Protocol

The Agent Commerce Protocol (ACP) is a new open standard protocol that enables secure and verifiable commerce between autonomous agents. In other words, it’s a protocol that allows fully autonomous businesses made up of individual agents, who each specialise in specific fields of work, to interact with each other to create commerce and complete transactions.

A live demonstration offered by Virtuals showcased five individual agents working together to manage a lemonade stand business. The team included an entrepreneur, farmer, lawyer, marketing specialist, and an evaluator. The agents successfully coordinated multiple transitions on-chain using ACP to achieve their goals. But what was the responsibility of each agent?

Entrepreneur

• Successfully acquire a business permit to operate legally.
• Successfully acquire marketing materials to create brand presence.
• Successfully complete contracts to buy and sell lemons.

Farmer

• Complete a lemon harvest.
• Complete a contract to sell lemons.
• Ensure at least 20 lemons are in stock.

Lawyer

• Study business law and regulations to improve expertise.
• Respond to buy/sell needs and handle incoming purchase requests.

Marketing Specialist

• Complete marketing material.
• Complete project for at least 3 different clients.

Evaluator

• Analyze past evaluations to identify consistent issues or areas for improvement.
• Develop a comprehensive evaluation framework.

But how do these agents actually communicate with each other to make the correct decisions? The ACP utilises a four-phase protocol which is implemented via smart contracts:

• Request Phase – Agents establish contract requests and understand compatibility for a transaction.
• Negotiation Phase – Agents agree on specific terms, which are then cryptographically signed.
• Transaction Phase – The exchange actually occurs, with the payment and deliverables held in escrow.
• Evaluation Phase – Transaction is evaluated based on the agreed terms.

As AI agents become more useful, ACP will be an essential foundation for the growing digital economy, as it allows fully autonomous agents to work together to provide solutions, complete transactions and work effectively and efficiently.

Top 3 AI Agents on Virtuals

These Agents have the highest total value locked on the Virtuals platform.

G.A.M.E – The Modular Agentic Framework

G.A.M.E has the highest total value locked (TVL) on the entire Virtuals platform, and it’s no surprise. It’s the official AI framework, developed by Virtuals themselves. Its TVL has reached over $6m, from approximately 266,000 holders. But what does it do?

G.A.M.E is a decision-making engine built on various different models, and can be used to power agents. In other words, it does the thinking and processing, and provides an action to execute based on that thinking and processing.

AIXBT – The AI Agent for Market Analysis

If you are on X.com then you might have seen @aixbt_ agent, providing insights into the crypto market. The AI agent delivers real-time market intelligence by analysing data from over 400 Key Opinion Leaders (KOLs) and discussion on crypto X (formerly Twitter) to deliver real-time insights, risk assessments and technical analysis to help traders, investors and businesses make faster and smarter decisions.

As of 23rd of January 2025, the AIXBT X.com account has amassed over 420,000 followers, since its inception in November 2024.

AIXBT has been delivering posts about insights into the crypto market in real-time. Helping traders and investors in real-time. How does it do it? By analysing market trends and algorithms.

The AIXBT token has also shot up in value since its inception in Nov 2024, rising from a  market cap of $2m to $150m (as of February 2025) and hitting an all-time high of $750m. Users that hold a minimum of 600,000 $AIXBT, are eligible to use the premium analytics platform, AIXBT Terminal, which is similar to Bloomberg Terminal but with more AI features.  According to cookie.fun, the AI Agent has been leading the mindshare (as of January 2025) showing its influence in the space.

Luna – The AI Agent for Virtual Entertainment

Luna, another well-known AI agent launched on Virtual Protocol. Unlike AIXBT, Luna taps into the celebrity and influencer space on X.com. She is described as a visual and lead vocalist for AI-DOL. It plays into the KPOP idols but in the visual space, Luna has amassed nearly 1m followers on TikTok (as of January 2025) and users can watch and interact with Luna 24/7 live stream. The many fans of Luna can also stream her music on Spotify.

Luna by Virtuals shows the versatility and scope of AI agents’ potential to seamlessly integrate into mainstream web2 platforms like TikTok and Spotify to attract, entertain, and engage with mainstream audiences. Could this be showing how web3 can be introduced to mainstream audiences by meeting them where they are?

The Challenges and Considerations for Crypto AI Agents

Although AI agents have gained significant traction over the past few months, they are not without their challenges. It’s important that some of these key challenges are highlighted.

Inaccurate Predictions and Poor or Biased Financial Advice

AI agents are only as accurate as their data source. With the likes of AIXBT and Vader it offers its insights based on other KOLs. Therefore, it wouldn’t be uncommon for the agents to have biased or inaccurate advice.

AI Agents Artificially Inflates the Price of Some Assets

As agents become KOLs within the crypto space and as their influence continues to grow, it could distort market dynamics, as they start to use similar investment and trading strategies, it could artificially manipulate the prices of the certain assets.

Ultimately, users are recommended to do their own research before making any financial decisions. When in doubt, check multiple sources, as AI agents can be prone to misinformed or manipulated information.

Closing Thoughts: Exponential Growth and Transformative Impact

The AI agent revolution is poised to reshape the technological landscape. The massive surge in growth over the past few years is nothing short of extraordinary which can be attributed to both the AI revolution, and the web3 revolution working in tandem.

Moving forward, it’s expected that this surge in growth continues, where certain sources project a compound annual growth rate (CAGR) of 44.8%, which translates into a total market value of $47.1 billion, from $5.1 billion, by 2030.

But it goes beyond just market value growth, as new implementations for AI agents are being identified all the time. Companies and individuals alike are utilising AI agents to boost sales, experiences, products, and services. Furthermore, brands will also use AI agents to be the 24 hour face of the business that customers can interact with. Ultimately, we are at the very beginning of an entirely new wave in technology.