The briefing is already in Telegram
Pilot sends the day's summary by itself: who is on shift, where coverage is short, yesterday's lateness, risks and recognition.
Smengo Pilot
Smengo Pilot is the AI layer of the Smengo operations system: it builds the schedule from labor rules and demand forecasts, answers shift questions from your data, sends morning Telegram briefings, screens resumes and assembles training. Anything that changes data happens only after your approval.
fill the August 14 gap with Marina
Marina K. is free on August 14 and has 128 hours in August — there's room before the cap. Here's a proposal:
Marina K. → August 14, floor, 09:00–18:00
What you can delegate
Pilot is not a separate app but a layer over your data: schedule, check-ins, hiring and the knowledge base. Here is an ordinary day.
Pilot sends the day's summary by itself: who is on shift, where coverage is short, yesterday's lateness, risks and recognition.
The generator assembles the schedule from labor rules, employee availability and the demand forecast. Edits in plain language: “take Marina off weekends”.
“Who can cover Friday?” The assistant answers from your data: 10 tools — 8 read the schedule, check-ins, the shift board and the budget, 2 more propose shifts.
Pilot proposes a shift — as a card. It reaches the schedule only after your Execute button.
AI screening: the candidate's strengths, weaknesses and questions for the call. The resume itself is never stored.
The onboarding board draws on your knowledge base, the quiz on the training text itself — you only edit and publish.
The problem
Generic “AI assistants” see neither shifts, nor check-ins, nor your rules — they answer in platitudes while the routine stays with the manager: build the schedule, answer “who works tomorrow?”, check lateness, dig up the policy clause. Every day, by hand, in messengers.
The schedule, check-ins, the shift board, candidates and documents already live in Smengo. Pilot is the AI layer above them: it drafts the schedule, answers with tools over live data, comes to you first with briefings and risks — and changes nothing without your button.
How it works
Say what you need — Pilot pulls the data, applies your permissions and returns a result you apply with one button.
Pilot runs inside Smengo: the schedule, check-ins, the shift board, hiring and the knowledge base are already linked. Zero setup, zero integrations.
Foundations ready
MODULES
DATA
Foundations ready
MODULES
DATA
Zero setup. Pilot runs on data already in Smengo: the schedule, check-ins, candidates, knowledge-base documents. No separate rollout, integrations or prompt engineer needed.
Full inventory
Check it against your requirements list: every line is a working product feature, not a roadmap item.
Generate a month's schedule
the solver honors labor rules, availability and the forecast — drafts up to 5 versions, button generation is free
Edit the schedule in plain language
“take Ivan off Tuesday” — a new draft version; nothing is written without Apply
Forecast demand by day
from revenue and traffic — Poster POS, CSV or manual input; shown in the schedule header and used by the generator
Understand why the schedule looks this way
a draft explanation: what got filled, which days stayed uncovered and why; viewing again is free
Ask “who works today?”
10 tools: 8 read the schedule, coverage, hours, availability, the shift board, lateness, demand and budget; 2 more propose shifts
Place a shift from the chat
only via an approval card — every action stays in the journal
Get a one-click briefing
the day summarized in plain words, up to 8 sentences right in the assistant chat; an empty day spends no credit
Break down a timesheet with AI
plan-vs-actual explained per employee and month: under- and overtime, days without a check-in; viewing again is free
Get a morning briefing
in Telegram by 08:00–08:30 in the organization timezone: shifts, coverage, yesterday's lateness, the shift board, risks and recognition; spends no credits
Catch risks early
a 28-day radar: lateness streaks, overtime, repeating coverage gaps, plan deviations and swap spikes
Spot employee overload
workload flags “watch” and “elevated” — managers only, nothing is sent to employees
Get “Week in numbers”
a weekly report by email and Telegram: planned vs actual hours, lateness, department deviations and risks
Lateness counts itself
bot check-in vs shift start with grace minutes — nothing is filled in by hand
Fill shifts without ring-arounds
offer waves in Telegram to matching people by department, position and load; those who asked for a shift go first; “Take it” in one tap
Play out “what if”
“+1 person” and “different coverage minimum” scenarios on a typical week — in analytics, spends no credits
Upload your company rules
PDF, DOCX and text, up to 50 documents — a semantic index builds itself and spends no credits
Ask your documents
the assistant's “From documents” mode: answers strictly from your fragments with a mandatory source-document reference
Let employees ask for themselves
the bot's “Ask the rules” button — an exact quote from your document with its source cited; free, up to 10 questions a day
Screen a candidate's resume
strengths, weaknesses and 5 call questions; contacts fill the card by themselves, the resume is not stored
Assemble onboarding for a newcomer
a 4–8 stage board from your knowledge base — in one generation
Build a quiz from a training doc
3–5 questions from the material; taken right in Telegram
Control
Orchestrator does not mean autopilot without a wheel: anything that changes data goes through a card with your button.
Marina K. → August 14, floor, 09:00–18:00
The card waits for your decision. Until the button is pressed, nothing in the schedule has changed.
Artem V. → August 16, kitchen, 08:00–17:00
At the moment of the press, conditions are re-checked: the cell is free, the labor budget isn't exceeded, the permission still holds.
Nikita L. → August 17, floor, 09:00–18:00
There is already an entry for this employee on that date.
While the card waited, the shift was filled by hand. The assistant doesn't overwrite someone else's work — the buttons stay, the call is yours.
Your data is not a textbook. Requests reach the model only at the moment of an operation, and models are not trained on your data. Resume texts are not stored at all after screening.
Pricing without surprises
No separate “AI subscription” and no unpredictable token bills — credits are included in your plan and every operation has a fixed price.
The server does the math. The operation price is set by the server, not the client — billing cannot be gamed. Your credit balance is visible in billing at any time.
Compare
A chatbot sticker and an AI layer on your data are different products, even if both spell AI.
AI as a layer, not a bolt-on Pilot lives inside the operations system and runs on data your team creates daily — that's why its answers match your schedule.
Schedule, bot, hiring and training — Pilot connects the modules rather than replacing them.
A board from applied to hired, AI resume screening, 5 interview questions.
Learn moreContacts behind a permission, documents with a 30-day badge, 100% history.
Learn moreAn AI plan of 4–8 stages, at-risk deadlines, 3–5 question quizzes.
Learn moreA month on one screen, 4 labour norms, AI generator, demand forecast.
Learn moreOne-tap check-ins, 28-day availability, swaps, a calendar subscription.
Learn moreAn approved swap becomes an open shift on its own; conflicts are blocked.
Learn moreHours from check-ins, XLSX/CSV in one click, 5 risk detectors, what-if.
Learn morerisks and coverage feed the next schedule ⟲risks and coverage feed the next scheduleThe AI layer of the Smengo operations system: a schedule generator, a data assistant, morning Telegram briefings, knowledge-base answers, AI resume screening and training generation. Not a separate product — it's included in Smengo plans.
No. Anything that changes data becomes a proposal card: a shift reaches the schedule only after the Execute button, and the decision is stored in the journal.
AI credits are included in plans: Start — 30, Team — 200, Business — 1000 per month. An operation costs 1 credit, a plain-language draft edit — 5, while button generation and the morning Telegram briefing are free. Failures refund the credit.
Only what the asker can see under their permissions: without payroll rights the assistant won't see rates. Models are not trained on your data, and resumes are not stored after screening.
It assembles a draft from labor rules, availability and the demand forecast; you edit it in plain language and publish yourself. Publishing is always your decision.
Russian, Ukrainian and English — like all of Smengo: the interface, the bot, briefings and assistant answers.
Nothing separate: Pilot runs on data already in Smengo. Briefings are a one-click subscription; the knowledge base fills up as you upload your documents.
ChatGPT doesn't know your schedule. Pilot answers with tools that read your organization's live data, respects access rights, and changes anything only with your approval.
Schedule, answers, briefings, hiring and training — AI is already included in the plans. Start with 14 free days.