In most companies, management accounting runs on people and files. As long as it works, nobody counts what delays and conflicting numbers actually cost
One pulls the exports, the other interprets them. Only a third finance professional could check the math, so the executive takes the numbers on faith
A cash flow summary hundreds of rows long, several versions of the file, manual reconciliation. By the time the report is ready, the situation has already changed
The owner, sales, procurement and finance all look at different exports. A dip in a channel or SKU is visible only to whoever goes looking for it
The lower layers make up the BI system: data, methodology, metrics and forecasting. The top layer is AI-CFO, which reads this data and works with it in place of a person
1C via OData, seller accounts on the Wildberries and Ozon marketplaces, bank statements, Central Bank of Russia exchange rates. Data loads into an analytical data warehouse on a schedule; raw data is kept separate from calculated data
A classifier for P&L and cash flow line items, allocation of overhead and unattributed cost of goods, mapping of items to channels. The rules are spelled out and open to review
P&L, cash flow, unit economics, inventory, receivables and payables, budgets, investments. Each view has its own drill-down, filters by legal entity and period, and a mobile version
A cash calendar with a minimum balance threshold, 30–90-day forecasts, what-if scenarios, department budgets, NPV and IRR for investment projects
Rules track deviations and gather them in the attention center. The assistant explains the cause, suggests an action with a dollar estimate and answers questions about the data. The language model is hosted inside your company’s infrastructure
Every metric is calculated from the same journal entries using the same rules. The owner, the CFO and department heads see the same numbers, not different exports. Any amount drills down to the account, analytical dimensions and source document in 1C
The BI system works on its own or as the data layer under AI-CFO and ERP
Click a channel to recalculate margin, average order and repeat sales
Each tab shows only what the system actually calculates and displays
A line-item tree from revenue to net profit, filters by channel and legal entity, period comparison. Click a row to open the accounts, analytical dimensions and the 1C document the amount came from
Unit cost on line #2 is up 6.3%. Cause: materials bought on Oct 14 above the standard price.
Lock the price in the supplier specification and add a check for purchase price deviations > 3%.
AI-CFO reads the BI system’s data marts directly. It spots a deviation before a person would, finds the document that caused it and suggests an action with its impact estimated in dollars. Ask a question by text or voice: the answer is based on the same numbers you see on screen
The assistant does not change data in your accounting systems: it suggests, and a person decides
Request demo accessThe logic comes from a live system. The numbers on this page are illustrative; the sequence of steps is real
A spreadsheet holds one person’s calculations. 1C reports show the accounting view. A dashboard displays whatever it’s fed. BI with AI-CFO calculates using a defined methodology and suggests action
| Criterion | Excel spreadsheet | 1C reports | Dashboard on exports | BI system + AI-CFO |
|---|---|---|---|---|
| Data source | A copy, often out of date | The accounting ledger | Whatever could be connected | 1C and marketplaces, data marts refreshed on a schedule |
| P&L methodology | In the file author’s head | Tax accounting logic | Whatever the client asked for | A classifier and rules open to review |
| Unit economics, ads, inventory | Three files that don’t match | Indirectly or not at all | Separate datasets | One SKU × channel model |
| Early warning on cash gaps | If someone remembers | Today’s balances | A chart, if someone set it up | Threshold, forecast, signal |
| Action | A meeting | A request to accounting | “Take a look at the dashboard” | A recommendation in dollars: accept or decline |
| AI | Public chatbot, data leaks | — | An assistant bolted onto the charts | Model on your servers, working from data marts |
A regular dashboard is still useful for one-off questions. BI with AI-CFO covers the everyday work of the owner and the CFO
See profit, control cash, plan ahead. All in one product, without separate spreadsheets in every department
It normally takes one finance professional to check another. The system follows rules that are open to inspection: mappings, the classifier, allocations and a link to the source document behind every amount
Journal entries, turnover and balances come in via OData; sales, commissions, advertising and stock come from marketplace seller accounts. The time of the last load is shown on screen
The P&L taxonomy, cost allocation rules, item-to-channel mapping and signal thresholds live in the settings, with a change log
Every amount in the tree drills down to the account, analytical dimensions and document. Reconciling with your analyst’s figures is routine, not a one-off project
Several legal entities on one screen; admin, finance, analyst and viewer roles; an activity log; a language model inside your security perimeter
In a live implementation, figures calculated from 1C journal entries matched the multi-year model kept by the company’s financial analyst over 39 months, with better than 99% accuracy.
Request demo access →to a board summary, instead of days spent gathering and explaining
of the numbers for the owner, finance, sales and procurement
cash gaps, margin dips and excess inventory show up in advance
every figure is traceable: mapping, classifier, 1C journal entry
Depth and impact depend on the quality of your 1C records, consistent use of analytical dimensions and full access to your marketplace accounts
The demo environment replicates a live system on anonymized data. You follow the same path as an owner: profit, cash, action
Profit tree, channels, EBITDA bridge. See where the company makes money
Balances, calendar, red zone. See where cash is fine and where the risk is
Signals with actions valued in dollars. Today’s priorities, not a list of notifications
Updated
A dashboard shows data marts that someone put together in advance. The BI system is a model of your finance domain: one methodology, every amount traceable to the journal entry, allocation rules, forecasts and scenarios. 1C reports show the accounting picture; the system builds the management picture
It works with the results. The assistant watches for deviations, looks for the cause in the documents, suggests an action with a dollar estimate and answers questions based on the data marts. The BI system does the math; AI-CFO explains and recommends
Usually two to three weeks: connecting 1C and the marketplaces, mapping line items, reconciling with your financial analyst’s figures. The outcome is a working system on your data, with success metrics agreed before we start
No. The model is deployed inside your infrastructure or on a dedicated server, and the data never leaves your perimeter. A cloud model is an option only if your security team separately approves it
The system will be the first to show it: transactions without a line item or department land in a posting quality report. The list of fixes goes to your accountants, and plan vs. actual starts to add up
It depends on the number of sources and legal entities, the depth of the methodology and the hosting option. After your request, we send you the scope of the system and an estimate for your accounting setup
