One ledger for the law and for the decision
A CFO lives with one contradictory requirement: the ledger must be accurate for the state and useful for the board, and in most systems those are two different sources. The report by plant is assembled in a spreadsheet and does not agree with the trial balance; the close takes twelve days; and the question of why the margin fell has three opinions and no data. NG One resolves that contradiction in the foundation: the dimension belongs to the ledger line, the parameter carries a validity date, and every posting references the rule version that produced it. Everything else follows from those three decisions — a report by plant that cannot diverge from the trial balance, a close that is a checklist rather than a race, and an answer to “why” that is read off the record.
- Dimensions in the core
- Soft and hard close
- Explain this posting
- Consolidation and intercompany
- Cash-flow forecasting
- Executive Cockpit
What takes a CFO's month
None of this is a question of effort. All of it follows from decisions taken before any of you sat down at that system.
The board report and the trial balance come from two sources
Analysis by cost centre, project or plant was added later — as an extra field, as a sprawling tree of analytical accounts, or as a parallel ledger. The report is therefore assembled in a spreadsheet, outside the system, slightly differently each month. When two numbers disagree, the discussion is not about the business but about which source is right — and that discussion never ends, because both sources have a case.
The close takes twelve days and half of it goes on reconciliation
Stock does not agree with the ledger, statements are matched by hand, advances hang unresolved, and FX differences are computed at the last minute. The close is therefore not a control but a race, and in that race materiality is decided by fatigue rather than by significance. By the time the numbers are final, the month they describe is six weeks old and is no longer a subject for decision but for history.
Cash flow is a spreadsheet maintained by hand
The cash-in projection lives in a file one person knows, runs on their laptop and changes when they get to it. Collection risk becomes visible once the customer has already stopped paying, not when their payment behaviour started to shift. The decision to defer a supplier payment is taken on a feel for liquidity, and it is usually right — until one time it is not.
Consolidating three entities is one evening and three exports
Each entity has its own export, elimination of internal relations is done by hand, and the intercompany balance is agreed by email with a colleague in another company. The result is as accurate as someone's available time and cannot be repeated without that same person — which means consolidation is in fact a personal skill rather than a process.
The rule changes in July and is applied in September
The system holds one active value for a rate or a mapping, so correcting a March document is computed under July's rules — and nobody notices until an inspection does. Deadlines are remembered, changes are tracked by reading, and the answer to “why was this posted this way” depends on whether someone recalls a setting from two years ago.
How NG One answers
The two things that make the difference are not features but decisions in the foundation: the dimension belongs to the line, and the parameter carries a validity date. Everything else stands on them.
Dimensions belong to the line, not to a layer beside it
Cost centre, carrier, project, plant and channel are part of the ledger line from the first posting. A project report leaves the same ledger as the trial balance and always agrees with it — there is no second source, so there is no argument about which is right. This is the one decision no system anywhere retrofits: a posting made without a dimension cannot acquire one without reposting history. So NG One posts dimensionally from the first line, not from the day the report is asked for.
A close that is a control, not a race
A period closes in two steps: a soft close stops routine work per module, a hard close locks the period. Reopening exists, but it requires approval and stays in the immutable journal — who, when and why. Reconciliation of stock, bank and sub-ledgers runs continuously, with a checklist of unreconciled items before the close. The business calendar and periods hold across every module: a closed period refuses a posting no matter which flow it comes from — sales, procurement, inventory or payroll.
“Explain this posting” as a property of the record
Every transaction references the rule version that produced it, so the answer to why something was posted a certain way is read from the record instead of reconstructed from memory. The answer is reproducible and identical for everyone who asks — you, the auditor or the inspector. This is a deterministic property, not an AI feature: effective-dated parameters mean the past is computed under the rules of the past, so correcting a March document two years later produces the entry March would have produced.
Consolidation, intercompany and agency mode
Multiple legal entities with elimination of internal relations, intercompany reconciliation, and the mode in which several companies are run in one system. Underneath sits a multitenant core with database-level data isolation and multi-currency; above it, a shared chart of accounts and shared dimensions — so a consolidated report is not assembled from three exports, it leaves the same ledger. The intercompany balance is reconciled in the system rather than by email with a colleague in another company, and it repeats without her.
A compliance calendar for legislation that changes mid-year
Rates, thresholds and tax mappings carry validity dates, so correcting a document from a past period uses that period's rules — a condition of accuracy, not a convenience. Filing deadlines come from the compliance rule pack, with the rule version and the source of the regulation in the trace. The screen with deadlines, regulatory changes and obligation status per legal entity sits in one place, and the statuses are live: what went out through SEF and was accepted or rejected, which record has been filed, and what is still outstanding.
An Executive Cockpit organised around decisions
Not a grid of charts but a chain: what changed → why (a breakdown of causes) → what follows if we do nothing → three decisions with what-if analysis and evidence. Every figure opens down to the line and the document. Cash-flow forecasting, collection risk per partner and anomaly detection run inside the same chain — and every flag carries the reason the record was singled out, because an alert without a reason is muted within a week.
A CFO's month
The same cycle the Atlas draws as record-to-report, seen from the chair of the person answerable for the number.
- Step 1
The document arrives
What can be prepared, prepares itself. What is left for a person is what requires judgement.
- OCR proposal with confidence per field type
- Duplicate check and a check that the period is open
- Link to the goods receipt or purchase order from context
- Automation KPI: what the system did on its own
- Step 2
Posting
The entry is assembled by the rule effective on the document date, and dimensions come from context — with no later allocation.
- Posting groups as configuration, not as code
- Posting preview: the full entry before confirmation
- Dimensions from the document context
- The rule version written onto the line
- Step 3
Controls and reconciliation
Reconciliation runs all month instead of being crammed into three nights before the close.
- Statement auto-matching by reference, amount and partner
- Open items, statements of open items, set-offs and cessions
- Stock-to-ledger reconciliation against inventory
- Maker-checker and SoD over financial actions
- Step 4
VAT and obligations
The obligation arises from the line under the mapping effective for that period — so the return is not a reconstruction of the month.
- Records in the EEO and EPP structure
- POPDV and PP PDV as submission-ready XML
- Versioned mapping by accounting period
- Deadlines from the compliance rule pack
- Step 5
Close
The close is a control with a checklist, not a race with a calendar. Reopening is a decision that leaves a trace.
- FX differences and revaluation of currency balances
- Soft close per module, then hard close of the period
- Checklist of unreconciled items
- Reopening with approval and a journal record
- Step 6
Decision
The report is not the goal but the input. The goal is a decision that can be explained and checked afterwards.
- Trial balance, account cards and journal with drill-down
- Reporting by dimension out of the same ledger
- Executive Cockpit: cause, consequence, three decisions
- Budget, controlling and cash flow by dimension
The spaces that carry this solution
A solution is neither a separate product nor a separate licence. It is the same system seen from one angle, and these business spaces carry most of the work this page describes.
Finance and compliance
Serbian statutory accounting, dimensions and VAT — in the core, not bolted on.
30 capabilitiesInsights, automation and AI
The command center for decisions — evidence behind every insight, a trail behind every automation.
30 capabilities
Questions about this solution
Scope, boundaries, and the rules this entry point works by.
How do I get a project report that agrees with the trial balance?
By having it leave the same ledger. The dimension — cost centre, project, plant, channel, vehicle, employee — is part of the ledger line from the first posting rather than an analytical layer beside it. The consequence is that a dimensional report and the trial balance cannot diverge, because they do not have two sources. That is a decision taken once, in the foundation, and it cannot be retrofitted: postings made without a dimension cannot acquire one without reposting history. It is therefore a question worth asking every vendor before signing, rather than after the first quarter.
How long does the close take and what does the system do on its own?
The honest answer: we do not know how long it takes at your company, and any number we wrote here would be marketing. What we can say is what makes a close longer — manual statement matching, stock that disagrees with the ledger, and advances left hanging — and what the system takes over: auto-matching by payment reference, amount and partner; continuous stock-to-ledger reconciliation; a checklist of unreconciled items before the close; and soft/hard close per module. How much of it the system did on its own is measured by the automation KPI in your tenant — with a number from your data, not from our brochure.
Does cash-flow forecasting work, and how far should I trust it?
It works, in two layers worth telling apart — that distinction is how you spot a vendor who knows what they are selling. The deterministic layer is exact from the first posting: open items by due date, a projection from contractual payment terms, and a payment proposal from due liabilities and available balance. That is not an estimate, it is arithmetic over your own documents. Above it runs the forecast: cash-in prediction from each customer's payment behaviour, collection risk scored before a document is approved, and detection of deviations in postings, duplicate invoices and unusual discounts. A forecast is only as good as the history beneath it — it starts on your migrated data and contractual terms, and with every month in the system it sees more precisely how your customers actually pay. That is true of every model on earth, and we say it out loud rather than promise you that a model knows something the data does not. Every forecast opens the lines it was built from.
How is consolidation of multiple legal entities handled?
Out of one system, not three exports. The multitenant core carries multiple legal entities, multi-currency and database-level data isolation; consolidation eliminates internal relations, and the intercompany balance is reconciled in the system instead of by email with a colleague in another company. What matters is why that is possible at all: the chart of accounts and the dimensions are a shared foundation, so a consolidated report needs no mapping between three different account plans and no reposting of history. Agency mode — several companies in one system, with separated data and shared work — is the same mechanism seen from an accounting firm's chair.
How do I know the AI is not posting behind my back?
Because it does not post. The AI reads documents, proposes postings, matches statements and prepares drafts — the result is always a proposal a person confirms, and approving, editing and rejecting are equal outcomes that all stay in the trace. In OCR, confidence is measured per field type and the system refuses to guess: an uncertain field stays empty rather than being filled with a probable value. The AI sees exactly what the user it acts for sees — the same permissions, the same tenant boundary — and every call enters the AI trail: input, model version, output, and who confirmed it. Unattended posting exists only where the rule is deterministic and explicitly enabled, and then it is rule-based automation, not a model's estimate. Above all of it sits six-layer authorisation with maker-checker and segregation of duties.
What does early enrolment bring for finance specifically?
The discount is the easy part: 40% off the solution price for enrolments by 1 September, against a 10% reservation advance. For a CFO the influence is worth more than the discount, and that is not a phrase. An ERP rollout has a handful of decisions taken once that carry the next ten years of reporting: the chart of accounts, the set of dimensions you will measure the business by, tax mappings, approval limits and segregation of duties. Early enrolment means those decisions run through your numbers while the scope is still being established — the analysis of your processes is the first step of the programme, not the last. Later, the same thing is a request added to a list after your first quarter of operation.
Test NG One against your own close
Book a working review for finance. We walk your chart of accounts, the dimensions you need, your closing flow and the reports your board asks for — on your numbers, in the system, with drill-down to the line and the document.