While some sellers argue over whether neural networks are useful, practitioners are actively handing routine marketplace dashboard operations over to autonomous AI agents. Leo Shevchenko, founder of the Merkatus academy, presented a real-world case of deploying digital assistants that manage stores on Wildberries and Ozon with virtually no manual human intervention.
Unlike rigid algorithmic scripts, AI agents built on modern language models have a deep grasp of context. Paired with marketplace APIs, the agents handle core operational blocks: instantly generating personalized replies to customer reviews while deftly addressing hidden negativity, monitoring direct competitors' price lists 24/7 and adjusting ad bids, and promptly updating listing SEO attributes to match shifting platform algorithms.
According to the entrepreneur, delegating routine work to neural networks saves a team up to 40 working hours per week. This lets the seller focus on what matters most—finding high-margin new products, watching unit economics, and negotiating with factories.
