AI implementation at an oil and gas manufacturing and trading company: accounts, 1C and sales in 7 sessions
An AI implementation case at an oil and gas company that manufactures and trades across three business lines. At the start, accounting was kept “for taxes”, the 1C accounting system sat on a home computer, and the client base lay dead in the CRM. In seven working sessions — about three and a half weeks — we built management accounts, moved 1C to a secure server and turned the client base into a sales dashboard.
Published · Breakdown by Anton Ro Baten
Case at a glance
- Company
- Manufacturing and trading company, oil and gas
- Scale
- 3 business lines
- Starting point
- Accounting “for taxes”, 1C on a home computer, a dead client base in the CRM
- Timeline
- 7 working sessions · about 3.5 weeks
- Status
- Implementation continues
- 7 sessionsabout 3.5 weeks — management accounts built and reconciled
- 5 / 5the owner’s and the manager’s roles verified
- 1 dayfor a client employee to build a dashboard on their own
- ₽20Ka month asked by contractors for bookkeeping support — done inside the environment instead
Point A: where they started
- Accounting was kept “for taxes” — there were no management accounts.
- The 1C accounting system sat on a home computer.
- The client base lay dead in the CRM.
What we did
Management accounts
Built and reconciled by the accountant. The verdict: “it all matched up”.
1C on a secure server
The database was moved from a home computer to a secure server with AI doing the work.
Client base → sales dashboard
The dead CRM base became a dashboard where AI acts as the head of sales. A client employee built the dashboard on their own in one working day.
Tender monitoring
Tender monitoring is running.
Role verification
The owner’s and the manager’s roles verified: 5 out of 5.
Point B: results
| Metric | Result |
|---|---|
| Management accounts | accounting “for taxes” → built and reconciled by the accountant |
| 1C | on a home computer → on a secure server |
| Client base | dead weight in the CRM → a dashboard with AI as head of sales |
| Tenders | monitoring running |
| Bookkeeping support | contractors asked ₽20,000 a month or ₽2,000–3,000 an hour → done inside the environment |
| Owner and manager roles | verified 5 out of 5 |
What didn’t work
- Freeing up the owner’s time depends on his decision on priorities between the three business lines — that decision is still ahead.
What’s next
- Decide on the priorities between the business lines and free up the owner’s time. The implementation continues.
Takeaways for your business
- Start with accounting: here management accounts were built and reconciled with the accountant in the first weeks — everything else rests on them.
- An employee builds a working dashboard in a day when AI has access to the data and a clear role.
- Check what you pay outsourcers for: a function contractors wanted ₽20K a month for was done inside the environment.
- An implementation runs into the owner’s decisions on priorities — make them in advance.
More cases
- Live AI caseCosmetics manufacturer on Wildberries & Ozon26automations in 6 weeks — 3 were planned
- Live AI caseCar rental for taxi drivers−20%payroll: ₽2.5M → ₽2.0M a month
- Scaling before the AI eraMarposadkabel×5revenue: ₽400M → ₽2B+