Autonomous Agentic Finance: AI Agents as On-Chain Market Makers and Liquidity Providers

Autonomous Agentic Finance: AI Agents as On-Chain Market Makers and Liquidity Providers This white paper explores the potential of large language model (LLM)-driven AI agents to support liquidity management in decentralised finance (DeFi), particularly within automated market makers (AMMs). It examines the challenges liquidity providers face, including loss-versus-rebalancing (LVR), impermanent loss, changing market volatility, transaction costs, maximal extractable value (MEV), and the complexity of managing concentrated liquidity positions. The paper introduces the proposed Agentic Liquidity Governance Framework (ALGF), a governance-first architecture that combines AI-assisted interpretation of market signals with deterministic risk controls, transaction validation, execution safeguards, and human oversight. It outlines a five-tier autonomy model, ranging from passive observation to open-ended autonomy, with clearly defined permissions, escalation procedures, audit logs, and criteria for increasing or reducing agent authority. In addition, the paper examines the potential effects of AI-managed liquidity on market depth, price discovery, liquidity concentration, and coordinated market behaviour. It addresses critical risks, including manipulated data, prompt injection, model errors, front-running, and accountability challenges. Finally, it proposes a shadow-mode evaluation protocol to assess agent decisions against passive and rule-based liquidity management strategies before any live deployment. The white paper presents a conceptual framework and research agenda rather than validated trading results. Its central contribution is a governance-oriented approach to exploring how AI agents could participate in on-chain liquidity management while maintaining transparency, enforceable limits, and clear operational accountability.