Я создал ИИ-ассистента, который ведет все мои дела (сэкономил 100к)

Summary

Mikhail Timochko describes using an AI assistant to centralize task management, calendar planning, and meeting follow-ups. His previous systems—including books and methods such as GTD, Scrum, and Essentialism, along with apps such as Todoist and Notion—were difficult to maintain. He would use them actively for about two weeks, stop updating them, and later spend hours rebuilding order from scattered notes and messages.

His proposed solution uses Claude with Notion and Google Calendar. He says the system organizes work and personal responsibilities into life areas, projects, and tasks; recommends a short list of daily priorities; updates records from voice messages; and captures meeting follow-ups. In his account, the main benefit is reducing the effort needed to maintain a task system and lowering the risk of forgotten commitments.

Workflow and Playbook

• Connect the tools: Link Claude to Notion, where tasks and projects are stored, and Google Calendar, where events are managed.
• Create a dedicated project: Set up a persistent Claude project, such as “My Life” or “My Tasks,” so instructions and context remain available across interactions. Timochko recommends separate AI projects for complex, recurring work that requires specialized context.
• Do an initial voice-based brain dump: Describe work and personal responsibilities, current projects, priorities, and open tasks. Claude then structures the information in Notion.
• Review the structure: Organize the system into life or business areas, projects within each area, and tasks within each project. Ask for more detail if the first pass omits important items.
• Use a daily rhythm:
  • In the morning, ask what needs attention that day; the assistant returns a prioritized plan.
  • In the evening, report what was completed and what new tasks appeared.
  • During the day, dictate new ideas, commitments, and calendar events.
• Run a weekly review: Spend about 30 minutes reviewing the full system, closing outdated projects, and adding new priorities.
• Capture meeting actions: Record calls and meetings, transcribe them, and give the transcript to the AI to extract tasks, owners, and deadlines. Timochko uses a voice recorder and the Bukvitsa transcription service; he says automatic Zoom transcription could be another option.
• Use the assistant during meetings: He describes asking Claude what outstanding issues should be discussed with his director. It returned 20 questions, 10 of which the participants had not remembered to raise.

Business and Operating Context

Timochko says he manages a content business and a subscription community, alongside a team and several products. His areas include Instagram, YouTube, Telegram, course development, platform work, live events, team management, and business strategy.

He also uses separate AI projects for tasks such as preparing presentations in his style, writing posts, developing course programs, and drafting YouTube descriptions.

He frames the assistant as a way to improve execution: track delegated work, record who owns each task and by when, schedule reminders to check progress, and retain decisions made in calls. The system also includes personal commitments, with the stated aim of keeping work and life obligations in one place.

Metrics and Claims

• Audience: 1 million Instagram followers, 100,000 YouTube subscribers, and 60,000 Telegram subscribers.
• Cashew Club: 5,500 members; the quoted subscription price is 2,990 rubles per month.
• Team: 15 people.
• Task system: Approximately 200 tasks across projects; the assistant typically identifies 3–4 main tasks for the day.
• Initial setup: He describes recording three voice messages of about 10 minutes each. He says a full brain dump could take 10–60 minutes, depending on the person.
• Ongoing maintenance: Roughly 30 minutes per week for a system-wide review.
• Optional Telegram bot setup: Estimated at 3–4 hours, requiring several programs and paid services.
• Savings and productivity estimates: He suggests a human assistant doing comparable work might cost 30,000–50,000 rubles per month. He estimates that productivity gains could be worth 100,000 rubles or more per month, potentially 1–3 million rubles for a large business. These are his estimates, not measured results stated in the subtitles.

Tools and Promotion

The setup centers on Claude, Notion, and Google Calendar. For voice input, Timochko mentions Claude’s mobile app and the separate Aqua Voice dictation app for computer use. For meeting transcription, he mentions Bukvitsa and Zoom transcription as alternatives.

Timochko also promotes his Cashew Club, which offers courses on content and neural networks. He says he uses Claude for content creation, editing, visuals, competitor analysis, customer development, finances, and some business-strategy decisions.

Presenter and Sources

• Presenter/source: Mikhail (“Misha”) Timochko, drawing on his personal workflow and estimates.
• Tools and services mentioned: Claude, Notion, Google Calendar, Aqua Voice, Bukvitsa, Zoom, and Cashew Club.