Longread translated automatically
big

As part of the first homework assignment for the AI media creation course, I developed a concept for a Telegram channel/community targeted at aspiring interior designers and architects.

Media Concept

As part of the first homework assignment for the AI Media Creation course, I developed a concept for a Telegram channel/community geared toward aspiring interior designers and architects.

The project unites emerging designers in a live space where they can gain practical knowledge, ask questions without fear of seeming foolish, and find support.

Media exists in a «channel + chat» format. The channel features articles, columns, and practical content, while the chat is dedicated to communication, Q&A, networking, and support. The name reflects the essence of the project: a draft is an honest stage of growth where mistakes and revisions are part of the process.The thematic focus remains the same: environmental and interior design; the narrow topic is practice for beginners. The three core pillars of the media are: specialized programs, project management, and inspiration. The project’s unique selling proposition also remains unchanged: it’s not a glossy magazine, but a draft; the chat is more important than the channel; the language is tailored for beginners, free of pretension.

Data Sources for Media

Media works with three types of data:1. Materials used to source topics and facts for articles: specialized Telegram channels (Design Mate, Fragments), reference websites (Behance, Pinterest, Archdaily), educational platforms (Geometrium, Bang Bang Education), professional literature, and podcasts.2. Data from chat and feedback—this includes subscriber questions, practical case studies, comments, and voting results. This is where topics for the «Draft Question» and «Best Story of the Week» sections are born. This is the most valuable source because it reflects the audience’s real-world challenges.3. Original projects: when a project exists, it is developed as a continuous narrative—brief, concept, revisions, budget. This is unique content that competitors do not have.

Data schema divided into human and machine processing

Human Processing:

— Topic selection and defining the narrative angle — Copywriting, fact-checking, and editing — Creating sketches, screenshots, and gathering references — Chat moderation and story curation — Final post assembly and publishing

Machine Processing:

— Sourcing and monitoring materials (RSS aggregators, channel tracking bots)

— Curating references by keyword (Pinterest, Behance)

— Generating draft imagery (Midjourney, DALL-E for memes and covers)

— Proofreading and style review (ChatGPT, LanguageTool)

— Engagement analysis (Telegram statistics, LiveDune)

Guidelines for Automated Services

Rules for ensuring machine processing maintains the channel’s tone of voice and concept.Guideline 1. Automation does not write the final text. AI can generate drafts, suggest phrasing, or check style, but the final voice of the channel remains with the author.Guideline 2. The visuals maintain the aesthetic of a «draft.» If image generation is used for memes or covers, the style must remain minimalistic, sketched, and devoid of gloss or 3D effects.Guideline 3. Automated source collection is strictly for monitoring. Bots may track new posts in competitor channels or new references, but the text is always manually refined. Direct copying is prohibited.Guideline 4. The chat is moderated by a human. Automatic filters can screen out spam, but selecting the «Best Stories of the Week» and responding to questions is done manually to maintain a lively conversation.Guideline 5. Transparency—if generated visuals are used in a post, this is not hidden, but it is also not emphasized. Aesthetic quality takes precedence over the «purity» of origin.

«Designer’s Draft» is a media platform where the human remains the author, and automation handles the routine tasks—monitoring, generating rough visuals, and proofreading text. This division allows the channel to retain its authentic voice while eliminating the need to expend energy on mechanical tasks. The core concept from the first assignment remains unchanged, but the data schema and guidelines make it robust and repeatable.

Черновик дизайнера
Project created at 08.11.2026
We use cookies to improve the operation of the HSE website and to enhance its usability. More detailed information on the use of cookies can be fou...
Show more