Automation
If someone does the same thing by hand every day, that is a script nobody has written yet. We take over the data transfers, exports, reconciliations and reports, and leave people the decisions rather than the copying.
Scope of work
point by point.
- Process audit and a saving calculation before we start
- Python scripts: processing, reconciliation, exports
- Scraping of websites and marketplaces, blocking handled
- API integrations between services that do not talk to each other
- n8n flows where the process has to be changed by someone other than a developer
- RPA where the system has no API
- Schedules, retries and failure notifications
- Logs and a panel showing the status of every run
- Documentation: how to restart it and what to do when it breaks
Situations
where this pays for itself.
Moving data between systems
Website enquiries into the CRM, CRM into 1C (Russian accounting system), 1C into the report. No Excel exports and no manual pasting.
6 hours a week per employee
Competitor price monitoring
Daily collection of prices and stock, a history of changes, and an alert when something falls outside your range.
1,200 items in 8 minutes
Reconciliation and month-end close
Payments matched, discrepancies found, the finished table ready on the first of the month.
Close in 1 day instead of 4
Reports that build themselves
Data from the ad accounts, the CRM and analytics in one dashboard with a morning email.
Report at 09:00 with no human involved
Four steps
from brief to handover.
- 01
Timing the process
We watch how the process runs today: how many steps, how many minutes, and where the mistakes usually happen.
- 02
Saving calculation
We price the process in hours and roubles. If automation does not pay for itself within six months, we say so plainly.
- 03
Pilot on one section
We run the script alongside the manual work and compare the results until they match.
- 04
Schedule and handover
We put it on a schedule, wire up alerts and show your team how to restart it and read the logs.
What we
build it with.
Three tiers.
The exact estimate follows the brief.
from ₽50,000
3–7 working days
- One process of up to 5 steps
- Script with a schedule and logs
- Failure notifications in Telegram
- Restart instructions
from ₽200,000
10–25 working days
- Up to 5 connected processes
- Integrations with CRM, 1C, marketplaces
- Panel showing run status
- Retries and failure handling
- n8n flows for edits without a developer
on request
25–60 working days
- Orchestration of dozens of processes
- Queues, priorities, failure isolation
- Metrics in Prometheus and Grafana
- Access control over the data
- A written on-call procedure
Prices are the lower bound. What pushes an estimate up is set out on the pricing page
About this
service.
01How do we tell whether a process is worth automating?
Multiply frequency by duration by the hourly cost of the person doing it. If the process takes more than 4 hours a week and repeats without any judgement calls, it will almost certainly pay for itself in 2–4 months. We work this out before the contract is signed and show you the arithmetic.
02What happens when a service changes its API or its markup?
The script fails and you are the first to know: a notification with the error text arrives in Telegram. A fix normally takes 1–3 hours. On a subscription, breakages like this are covered by the support hours.
03Is scraping legal?
Collecting open data is legal; collecting closed or personal data is not. We do not get around authentication, we do not collect personal data, and we respect robots.txt and rate limits. If a task falls outside that, we turn it down and explain why.
04Will we be able to change the logic ourselves?
Where the logic changes often, we build the flow in n8n—a visual editor a manager can edit. Where speed and heavy processing matter, we write Python and hand over commented code.
Get an estimate:
Automation
The brief takes 5–7 minutes. In working hours we reply within two hours, and the estimate is free.