US-based startup Axal is developing automated crypto trading tools powered by AI agents, targeting non-technical users interested in crypto but not experienced traders or DeFi experts.
Axal’s first product, Autopilot, enables users to automate crypto and stablecoin trades by setting preferences such as risk tolerance and other restrictions. The tool is designed for ease of use, making it ideal for non-technical crypto fans who are interested in the space but aren’t day traders or DeFi experts. Autopilot currently operates using Ethereum Layer2 (L2) chains—Base, Optimism, and Arbitrum—with upcoming plans to support the Solana blockchain.
Users can fund their Autopilot accounts either through crypto wallets or via email, including traditional payment options like Venmo. Once preferences are set, AI agents manage tasks including trading cryptocurrencies (BTC, ETH, memecoins, and more), collecting stablecoin yields, and utilizing DeFi tools; users can also create custom token indexes for the AI to trade on their behalf. According to Axal, there’s a wide range of functionality with Autopilot, and it seems like there are enough parameters that can be set to ensure there will be no major surprises.
Acknowledging the risk of AI “hallucinations” (incorrect outputs), Axal leverages the transparency of blockchain to allow users to verify that AI actions are executed as instructed. “You can literally look onchain that this thing that you wanted this agent to do, it actually did,” Axal Head of Strategy and Operations Ari Santos said. Technically, Axal’s system incorporates zero-knowledge co-processors and Optimistic oracles, providing mechanisms to verify task completion and accuracy.
To prevent persistent errors, models that provide inaccurate results have collateral at stake and can be penalized (“slashed”), directly discouraging AI from making mistakes. “If a model hallucinates, it still has collateral at stake and it gets slashed. Whoever’s running the model, or the model itself, if it has very low accuracy, it’s just taking that risk,” Axal founder Ashlan Ahmed explained.
Axal’s roadmap includes additional AI tools for everyday automation, such as ordering food based on general requests or booking travel according to user input. The company believes that the future success of AI agents depends on verifiable completion of tasks, fostering user trust in agent-driven automations. “There’s a lot of different people in the agentic space,” Axal Head of Growth Jacob Kozhipatt acknowledged. “What we believe is that in order for the agentic feature to happen, people need some form of guarantee that it’ll actually be done the way that you asked to be done.”