One of the goals of 'AI for Science' is to discover customized materials through real-world experiments. Pioneering advances have been made in computational predictions and the automation of materials synthesis1-7. Yet most materials experimentation remains constrained to using unimodal active learning approaches, relying on a single data stream. The potential of artificial intelligence to interpret experimental complexity remains largely untapped8,9. Here we present Copilot for Real-world Experimental Scientists (CRESt), a platform that integrates large multimodal models (incorporating chemical compositions, text embeddings and microstructural images) with knowledge-assisted Bayesian optimization and robotic automation. CRESt uses knowledge-embedding-based search space reduction and adaptive exploration-exploitation strategy to accelerate materials design, high-throughput synthesis and characterization, and electrochemical performance optimization. CRESt enables monitoring with cameras and the generation of vision-language-model-driven hypotheses to diagnose and correct experimental anomalies. Applied to electrochemical formate oxidation, CRESt explored more than 900 catalyst chemistries and 3,500 electrochemical tests within 3 months, identifying a state-of-the-art catalyst in the octonary chemical space (Pd-Pt-Cu-Au-Ir-Ce-Nb-Cr) that exhibits a 9.3-fold improvement in cost-specific performance.

The copper-catalysed functionalization of aryl halides is one of the most preferred methods for forming carbon-carbon and carbon-heteroatom bonds1. Yet the redox behaviour of the copper species in the catalytic cycle remains poorly understood and a subject of debate2. We report experimental and theoretical mechanistic investigations into the reaction of a well-defined Cu(I) complex with an electron-poor aryl iodide, which leads to the formation of an isolable Cu(III)-aryl complex that subsequently reductively eliminates to form a C(sp2)-CF3 bond. Our integrated experimental and theoretical findings indicate that the process proceeds through a Cu(I)/Cu(III)/Cu(II)/Cu(III)/Cu(I) redox sequence. By controlling the temperature, we managed to interrupt this sequence and capture the reactivity of the copper species through various spectroscopic methods, enabling in-depth mechanistic analysis. These findings shed light on the intricate behaviour of copper species and challenge the traditional mechanistic proposal for the reaction of Cu(I) with aryl iodide, thus providing fresh perspectives into the mechanistic aspect of the copper-catalysed coupling reactions.