How it works
Your question goes through several steps before the answer comes back. Understanding this helps: you'll know why some questions get a precise answer and others don't.
01 — how it works
What happens when you send a question
You write a question
Just type naturally in the chat. No commands or special format needed. The bot understands free-form text — write the way you think.
"How do I delegate tasks properly without having to redo everything myself?"
Search in the knowledge base
The bot searches a selected base of Maxim's material — case discussions, talks and publications — not the internet. These are specific words from specific breakdowns.
Answer generation
The retrieved material gets formatted: noise removed, structure added. The tone is direct and partner-like — the way it sounds in a session. If there's a relevant video or post by Maxim, the bot will include a link.
Coaching question
A detailed answer comes with one question from Maxim. Not decoration — a working tool: it surfaces contradictions, shifts focus, or drives toward a concrete action. It doesn't follow every answer: on reference and clarifying queries it is suppressed on purpose.
"What would have to be different in your situation for this move to work?"
02 — entry
What happens right after the button
The first screen isn't a greeting — it's a gate. The product is in closed alpha, and the same screen asks for consent: during alpha your questions and the bot's answers are stored and analysed, which is how the product gets sharper. Until you agree, the bot answers nothing.
After "Yes" you get the TRIAL level and seven questions straight away, while an administrator confirms access. Language — Russian or English — is chosen on the same screen.
If you arrive after the diagnostic
The diagnostic gives you a coordinate: what stage the business is at, what fits that stage and what would hurt. You can work through it right here — with your own questions about your own stage. The coordinate sets the topic of the conversation; today it is not fed into the answer itself.
03 — access
Access levels
Eight levels live in production; the three below are the ones almost everyone passes through. What you get depends on your participation in Maxim's programs.
Entry after consent
Where anyone arriving from the site lands
- ·Answers up to 600 words
- ·Material name under the answer — no links
- ·Context: the last 5 exchanges
Basic access
Answers up to 600 words
- ·Questions on open program materials
- ·Context: 3 exchanges within a single day
- ·Related videos and posts — no links to lessons
Full access
For participants of Maxim's courses and programs
- ·No question limit, answers up to 800 words
- ·The bot remembers: 10 exchanges plus a session summary
- ·A coaching question where it fits
- ·Sources with date, timecode and link
Personal sessions and consultations are closed by level: only MENTEE and TRUSTED see them. Arriving from the site, you work with open materials, the author's base, partner programs and "Practicum Business".
04 — access without programs
How to get access if you're not a program client
There is a path, and it isn't manual: a robot raises your level on activity — at least three questions across at least two different days. If the alpha is full, you join a waiting list. The manual path is open too: the "How do I become a client?" button leads to a direct contact.
The bot takes no money: no cards, no subscriptions, no auto-renewal — there are no payment modules in it.
05 — commands
Controlling the bot
The bot's menu holds fourteen commands. The most visible is /style: three tones to choose from.
/start The beginning: the alpha screen and consent. /help Help — short or detailed. /style Tone of the answer: standard, tougher, gentler. /lang Answer language: Russian or English. /myinfo Your access level, stats and settings. /new_chat New topic: the bot forgets the previous conversation. /context The full history of your questions. /toc The table of contents of your program. /bookmarks Your saved answers. /insight The insights store as an entry point. /tracker Your personal goal tracker. /notifications Reminders: on or off. /pause Pause the conversation with the bot. /onboarding Go through the introduction again. 06 — the conversation
The conversation stays yours
You can see what the bot remembers
The /context command shows what part of the conversation the bot currently holds. Memory is the last few exchanges, not the whole correspondence; how many depends on your access level.
Start over with one command
/new_chat clears the context: your next question opens a new session, and the earlier conversation won't affect it. On the free level neither the history nor the manual reset works — there the context clears itself, every day.
Keep what's yours
An answer you like is saved with the button under it, and /bookmarks opens your own library: show it in full, add a note, set a tag, share, delete. Nobody else's bookmark can be opened by any identifier — a per-user filter sits on every read and every write. That's how the storage is built, not a promise.
07 — data
Data and privacy
What is stored
The full text of your question and the answer, a snapshot of the pipeline run, your name, e-mail and segment profile. The consent screen says so — once, at entry.
What the system learns from — and what it doesn't
Your conversations never go into model training — the model does not retrain itself on your questions, there is no such mechanism here. Something else happens, and it runs through a person: you rate an answer, a poor rating enters a review queue, and a good question is added by hand to the set answers are checked against.
Other people's material never reaches you
The per-client filter sits at the retrieval step, not after it: other people's material physically cannot enter your answer. Personal consultations are isolated strictly.
What isn't there yet
There is no "delete my data" command for users — deletion is done by an administrator on request. The bot also has no separate privacy policy yet. We say it here so you don't find it out later.
08 — limits
What the bot doesn't do
- —Text only. Voice messages aren't processed — the bot says so plainly. Photos, documents and stickers get no reply.
- —It doesn't answer from general knowledge. No material — the bot says it didn't find any, logs the question and passes it to Maxim.
- —It doesn't advise off-topic: weather, medicine, law, code, politics — a direct refusal.
- —Answer length is capped by access level — from 600 words at entry to 800 for clients.
09 — tips
How to ask for precise answers
Be specific
The more precise the situation, the more precise the answer. "How do I delegate?" works worse than "How do I hand sales to a manager when I'm afraid quality will drop?".
Talk as you would to a mentor
Don't hunt for the "right" wording. Describe the situation in your own words — the way you'd tell Maxim in a session. The bot understands natural language.
Go deeper
Got an answer — ask a follow-up about it. For clients the bot remembers the conversation, so you can unfold a dialogue instead of starting over.
Ready to try?
The first screen will ask for your consent — the product is in closed alpha. After that, seven questions straight away.