Research Disclaimer
Crypto.com Research and Insights disclaimer for research reports
Executive Summary
- An AI agent is a system that can autonomously operate or perform tasks on behalf of a user or another programme. Globally, the AI agents market is expected to grow from US$5.1 billion in 2024 to $47.1 billion by 2030, representing a 45% CAGR, according to MarketsandMarkets™.
- Web3 provides a good environment for AI agents. Blockchain eliminates intermediaries from transactions through automatic execution of smart contracts and enables AI agents to process transactions on behalf of users.
- The current landscape of AI agents in crypto include Agent Creation, Trading, Smart Wallets/Payments, Gaming, Art/NFT, and Security. Amongst the listed tokens in the AI agents landscape, Artificial Superintelligence Alliance (FET) has the largest market capitalisation with $4.2 billion (at the time of writing). It is an open platform to build innovative AI apps and services, including AI agents.
- Crypto.com recently released an AI Agent SDK, which aims to empower developers — and ultimately end-users — to interact with the Cronos blockchain and other Crypto.com services by leveraging AI tools as an advanced intermediary. At the time of writing, the current version is able to handle functions like calling chain data (e.g., balance enquiry), wallet management (transfer functions), and basic smart contract interactions (swap). New capabilities will gradually be released.
- Application of AI agents in crypto is in the early stages, but we believe that in the future, AI agents’ development in crypto could evolve into a multi-agent network and intent-based interactions — where users can type in any commands, and multiple agents in the backend would work together to provide verifiable, reliable, and secure outcomes to users.
1. Introduction
Artificial intelligence (AI) agents are not a new concept, but their capabilities have widely expanded with the rise of generative AI and large language models (LLMs).
An AI agent is a system that can autonomously operate or perform tasks on behalf of a user or another programme. AI agents can perceive their environment and take autonomous actions to achieve goals set by humans. They may also improve their performance through learning or acquiring knowledge.
Traditional chatbots and trading bots generally follow a predefined script (for example, if condition A is met, action B is executed). On the other hand, AI agents like AI chatbots can analyse various indicators, large datasets, and market sentiment, in addition to integrating with various APIs, to make decisions accordingly. In other words, AI agents can understand and reason. They also run on self-reinforcing algorithms, which means they learn over time and have the ability to refine outputs accordingly.
Globally, the AI agents market is expected to grow from US$5.1 billion in 2024 to $47.1 billion by 2030, representing a 45% compound annual growth rate (CAGR), according to MarketsandMarkets™. Growth is expected to be driven by demand for automation and efficiency across industries, including healthcare, finance, and customer service. In particular, multi-agent systems are poised to see rapid growth given the complexity and dynamic nature of most industries. Some major companies in the market include Google, IBM, OpenAI, and Amazon Web Services (AWS).
This report looks into AI agents and their applications in the cryptocurrency realm.
1.1 AI in Web3
Though AI in Web3 is still in its early stages of development, Web3 provides a good environment for AI to develop. In our previous report, we discussed potential developments where AI and blockchain can work together to solve issues in both areas and create innovative use cases. For example, with blockchain’s immutable digital record, we can potentially understand the framework behind AI models and the data they are using. Additionally, blockchain can help reduce costs associated with training and inferencing AI models by utilising the computing power of many machines. Moreover, zero-knowledge machine learning (zkML) can be used to verify AI computations on-chain or off-chain without revealing sensitive information.
Particularly, with the support of smart contracts that can be automatically executed, intermediaries are eliminated from transactions, thereby removing the required trust of third parties while increasing efficiencies across transactions. This drives the development of AI agents in Web3 and enables them to perform on-chain interactions with smart contracts and users’ wallets, enhancing the overall functionality of decentralised applications (dapps) and improving user experiences.
Blockchain technology opens up the possibility for AI agents to act on behalf of users, connecting to wallets, owning assets, and processing transactions. This enables AI to not only suggest actions and provide responses, but also to execute them from start to end.
1.2 Mechanism of Web3 AI Agents
In Web3, AI agents mostly host computations off-chain for cost efficiency, while agent decisions are executed on-chain.
The general workflow is described below.
- Users set their goals (intents); for example, purchase 1 ETH with ARB.
- The intent is processed by the AI model, which will request more information from the user if needed (e.g., timing and price of the purchase).
- Depending on the actions required, one or more AI agents are involved, communicating with various resources, such as databases, APIs, or external services, to gather the information needed (e.g., obtain the price of ETH, swap tokens to get more USDC, etc.).
- The LLM generates a response to address the user’s enquiries, who can either approve or provide input to refine the outcome.
- AI agents aid in execution with appropriate wallet integrations and authorisation.
2. Web3 AI Agent Landscape
Below are a few of the listed tokens in the AI agent space.
Artificial Superintelligence Alliance (FET), which has the largest market capitalisation ($4.2 billion) at the time of writing, is an open platform to build innovative AI apps and services, including AI agents.
Below are examples of current and developing use cases of AI agents in crypto.
| Categories | Description | Examples |
|---|---|---|
| Agent Creation | Creators can create/co-create, deploy, and monetise AI agents | Fetch.ai, Phala Network |
| Trading | Monitor a range of technical and non-technical indicators to build and execute trading strategiesOffer insights and probabilities to prediction market outcomes Portfolio management (e.g., liquidity management, optimise borrowing rates) | Autonolas (Olas Predict), Spectral Labs, Noya, Aperture Finance |
| Smart Wallet/ Payment | Non-crypto use cases (for example, to conduct payment transfers or make travel bookings) | Crypto.com’s AI Agent SDK, Skyfire |
| Gaming | AI agents deployed as characters to interact with users and create a changing gaming environment | Parallel’s Colony (Wayfinder) |
| Art/NFT | Creation of NFTs with interactive capabilities and characters through AI | Alethea.ai, NFPrompt |
| Security | Detect suspicious behaviour and find anomalies (e.g., in smart contracts) to protect users from scams | 0x0.ai |
| Categories | Agent Creation |
|---|---|
| Description | Creators can create/co-create, deploy, and monetise AI agents |
| Examples | Fetch.ai, Phala Network |
| Categories | Trading |
| Description | Monitor a range of technical and non-technical indicators to build and execute trading strategiesOffer insights and probabilities to prediction market outcomes Portfolio management (e.g., liquidity management, optimise borrowing rates) |
| Examples | Autonolas (Olas Predict), Spectral Labs, Noya, Aperture Finance |
| Categories | Smart Wallet/ Payment |
| Description | Non-crypto use cases (for example, to conduct payment transfers or make travel bookings) |
| Examples | Crypto.com’s AI Agent SDK, Skyfire |
| Categories | Gaming |
| Description | AI agents deployed as characters to interact with users and create a changing gaming environment |
| Examples | Parallel’s Colony (Wayfinder) |
| Categories | Art/NFT |
| Description | Creation of NFTs with interactive capabilities and characters through AI |
| Examples | Alethea.ai, NFPrompt |
| Categories | Security |
| Description | Detect suspicious behaviour and find anomalies (e.g., in smart contracts) to protect users from scams |
| Examples | 0x0.ai |
2.1 Agent Creation
Anyone in the community can participate in creating AI agents. For example, developers and AI users can work together to deploy an AI agent depending on their specific needs. This is enhanced by open-source frameworks, which enable collaborative refinement of AI agents in the community. Moreover, this can be a potential business model for creators to monetise their agents.
An example is Fetch.ai’s Agentverse, which is a cloud-based service for creating, testing, and deploying AI agents. It provides tools and prebuilt templates to facilitate the process of building and training the agents. In addition, it acts as a search engine to discover agents registered on the network.
Screenshot of Agentverse from Fetch.ai docs
2.2 Trading
Currently, one of the most common use cases for AI agents in crypto is to facilitate trading. AI agents can have the ability to incorporate technical indicators and retrieve data from external sources to build trading strategies. For example, users can type in commands like “Purchase BTC if the US Fed cuts interest rates by 50 bps in September.”
One example is Spectral Labs’s Syntax, which is an LLM to convert natural language into deployable Solidity code. The protocol is working on integration with pricing oracles and APIs (e.g., DexScreener, Tradingview), DeFi applications, and social media sites (e.g., X, Discord, Slack, Telegram). This would enable users to specify various trading and non-trading indicators in their strategies, and use it to explore on-chain data or back-test transactions.
2.3 Smart Wallets/Payments
In addition, AI agents can be equipped with their own wallets (separate from user wallets) to automatically process wallet transactions on behalf of the user. Furthermore, users can transfer funds to this separate wallet and authorise the agent to handle the funds.
Use cases are not limited to cryptocurrency. With the convenience brought by this smart wallet integration, automation can be brought to daily life payment use cases. For instance, there’s potential to use AI to check out automatically in online purchases or make travel bookings, etc. Skyfire has recently launched its payment network for AI agents, where each agent is assigned a digital wallet, and businesses can transfer a set amount of funds into the wallet to be used automatically. During the beta test, companies have been using the feature to pay contract workers autonomously and conduct month-end wire payments.
2.4 GameFi and SocialFi
We also see potential applications of AI agents in GameFi and SocialFi. Traditionally, games have predefined scenarios and environments designed by the creators. However, AI agents can potentially augment gaming experiences by creating an environment that changes based on the interactions with users. In addition, AI agents can also be deployed as characters within the game to create new possibilities and user experiences.
One example is Parallel’s Colony, a survival simulation game currently under development on Solana. In the game, AI agents and users work together to complete tasks, but the agents are not under direct control of the humans; instead, users make suggestions to the agent, which will in turn make their own decisions. The addition of AI agents creates more possibilities and variations in gaming, which can potentially aid in user retention.
Screenshots of Colony from X (Alpha Access in Q1 2025)
Within SocialFi, AI agents can be deployed to optimise social interactions. On the user end, SocialFi platforms can leverage AI agents to recommend personalised content to users through learning user behaviour, which can enhance content quality and user experience. On the creators’ end, AI agents can aid in content generation and monetisation. Agents can perform data analysis on social media sites to understand sentiment and trends, which enables creators to optimise content (e.g., type of content, target audience, etc).
2.5 Art/NFT
AI agents are also deployed in the NFT space and in art creation. Users can provide prompts (e.g., videos, images, text), and AI helps to create NFTs. In addition, AI agents can be used to add personalities and characters to NFTs and foster interactions.
An example is Alethea.ai, which enables the creation of smart interactive avatars through AI, known as iNFTs (ERC-721 assets fused with intelligence). For example, the avatars have personalities and are able to respond independently through generative AI. The company created the world’s first iNFT, Alice, which was sold at a Sotheby’s auction for $478,800 in 2021.
Screenshots of Alethea.ai’s AI engine
2.6 Security/Privacy
AI agents can be used to ensure security and privacy, as well. Using machine learning algorithms, agents can learn to detect anomalies or suspicious transactions to protect users from potential scams or fraud.
0x0.ai is an example project that focuses on the development of AI-based safety tools. This includes an AI smart contract auditor (currently under beta testing) that identifies potential vulnerabilities in smart contracts to ensure security. The project also has a privacy mixer, which pools multiple transactions and redistributes them to the intended recipients to make the transactions more private.
Screenshot of 0x0.ai’s Smart Contract Auditor (beta version)
3. Crypto.com’s AI Agent SDK
Crypto.com recently released an AI Agent SDK, offered as a PyPi and NPM package. The SDK can process natural language, map it to blockchain functions and execute instructions directly. It aims to empower developers and, ultimately, end-users to interact with the Cronos blockchain and other Crypto.com services by leveraging AI tools as an advanced intermediary.
At the time of writing, the latest version is able to handle functions including:
- Calling chain data (e.g., balance enquiry)
- Wallet management (create and transfer funds)
- Smart contract interactions (swap)
New capabilities are being developed and will be gradually released, including enabling a wider range of smart contract interactions (eg. ERC-20/ERC-721) and extending the wallet coverage to support more providers.
With the SDK as one of the building blocks, developers can further develop applications leveraging AI and blockchain, including:
| Use Cases | Description |
|---|---|
| Automated Portfolio Management | Continuously scans and generates insights based on on-chain data, and automatically executes trades when certain conditions are met. |
| Blockchain Analytics and Reporting Tool | Acts as a chatbot to transform user questions (e.g., “What is the market capitalisation of Ethereum now?”) into actionable blockchain queries, generating valuable insights on market trends and implications. |
| DAO Management | Translate governance proposals or community feedback into binding smart contracts and agents can help to manage the execution |
| Content Monetisation | Creators can use this AI agent-empowered wallet to receive tips and payments. |
| General Commercials | Embeds crypto payments into shopping, travel services, subscriptions, and more. For example, AI agents can help to browse through shops to find discounts and initiate payments, or execute travel plans to manage bookings via crypto wallets |
| Use Cases | Automated Portfolio Management |
|---|---|
| Description | Continuously scans and generates insights based on on-chain data, and automatically executes trades when certain conditions are met. |
| Use Cases | Blockchain Analytics and Reporting Tool |
| Description | Acts as a chatbot to transform user questions (e.g., “What is the market capitalisation of Ethereum now?”) into actionable blockchain queries, generating valuable insights on market trends and implications. |
| Use Cases | DAO Management |
| Description | Translate governance proposals or community feedback into binding smart contracts and agents can help to manage the execution |
| Use Cases | Content Monetisation |
| Description | Creators can use this AI agent-empowered wallet to receive tips and payments. |
| Use Cases | General Commercials |
| Description | Embeds crypto payments into shopping, travel services, subscriptions, and more. For example, AI agents can help to browse through shops to find discounts and initiate payments, or execute travel plans to manage bookings via crypto wallets |
4. Conclusion
The application of AI agents in crypto is currently in the early stages — the applications mentioned above are still gradually being developed and refined.
However, AI agents can be seen as a significant integration into crypto, as they can potentially act as a catalyst to onboard the next wave of retail users. While users may have an impression that crypto is difficult to navigate, with AI agents, they simply have to type in text commands and interact with the agents to execute tasks. In other words, AI agents enable users to bypass the hassle of managing wallets and navigating bridges and various sites.
As AI agent developments continue, we think agents can further improve user experience by enabling intent-based transactions — where users can type in any commands and agents are able to interpret and execute them. In addition to enabling crypto transactions, AI agents empowered with smart wallet capabilities also open up use cases in our everyday lives — for example, AI-enabled purchases and payments, etc.
We believe that, in the future, AI agents’ development in crypto could evolve into a multi-agent network with intent-based interactions to provide verifiable, reliable, and secure outcomes to users.
Read the full report: AI Agent in Crypto
Authors
Crypto.com Research and Insights team
Get the latest market, DeFi & NFT updates delivered to your inbox:
Be the first to hear about new insights:
