Content production — AI-assisted expert articles for leading media

A step-by-step process for producing expert articles for business media (RBC, Sostav, VC.ru) with generative AI, which made text production four times faster.

The goal was to raise brand awareness among business owners, CTOs and marketers. We needed to strengthen the company’s expert profile in the market, reach a new audience and build the kind of trust that generates a steady flow of inbound enquiries.

Industry: public relations, media, content marketing
Technologies: large language models, Gemini, Grok
Timeline: a continuous process (preparing one article went down from 10–12 to 2–3 hours)

The key requirement was adapting the narrative style strictly to each outlet’s audience — business leaders, the advertising market or the community of technology entrepreneurs. To make the text sound natural we used multi-step prompting, which removes typical AI clichés and integrates the call to action in line with advertising law. The whole process is a close partnership between AI and people: the algorithm takes over the routine first draft, while the editor owns the strategy, deep fact-checking and the final meaning.

Article preparation is split into seven stages. After the topics are chosen and internal facts collected, the model produces a base draft without promotional wording. A series of refining prompts sets the author’s voice and carefully weaves in the company’s experience. The data is then fact-checked rigorously to rule out model errors. Illustrations are generated by visual models in the format of the specific outlet, and at the final stage a human editor checks originality, logic and compliance with the publication’s requirements.

The process cut the time to prepare an in-depth expert article from ten hours to two or three without losing quality. The company published expert materials on the country’s key business platforms and reached a new business audience. The result is a transparent, scalable content production system in which AI accelerates the work many times over, while people keep the decisive role in strategy and in verifying what is published.