Inside SAP, the numbers are right. Taken out of SAP — into a data platform, BW, Power BI or a spreadsheet — they lose the rules that make them right. We write those rules down once, as a tested semantic model you own: the foundation every report and AI answer builds on.
SAP has the right answer. But ask the same simple question outside SAP and you get several. Only one matches SAP and the books — and it takes your one SAP expert three days to produce.
A semantic model is the foundation that fixes this: SAP’s rules, written down once and tested, so the right answer comes out everywhere — in every report, every dashboard and every AI answer.
“What did we sell to our top customer last quarter?”
With ECC mainstream maintenance ending in 2027 and AI on every board agenda, a semantic model of your SAP data is needed now — as the foundation for your reports, your AI and your move to S/4HANA. Which of these sounds like your SAP landscape today?
Reports are SAP downloads stitched together in Excel — and one “Excel guy” knows which lookups and macros make the numbers work.
With the semantic model: their logic written down once and tested, so the spreadsheets can retire one by one.
New BW reports take months, so people pull their own extracts. And BW 7.5 mainstream maintenance ends in 2027.
With the semantic model: BW’s business logic documented, so it survives any move — and the side extracts stop.
All the data is there, and fast. But reports and AI read tables, not rules — and the numbers don’t match the close.
With the semantic model: Silver, Business Vault and Gold on top of your Bronze — SAP’s rules, applied once.
Years of Z-tables, custom fields and user exits. Standard content and SAP’s own data products don’t know them.
With the semantic model: your custom logic mapped next to SAP standard — and carried to S/4HANA.
MRP not switched on, planning in spreadsheets or side tools. SAP holds only part of the truth — and nobody is sure which part.
With the semantic model: a clear map of where SAP’s data ends and the side tools begin — the base for switching modules on properly.
Warehouse, CRM, e-commerce, EDI and planning tools feed and read SAP — each with its own keys and its own version of a customer or product.
With the semantic model: SAP’s keys as the backbone: one customer, one product, one order — across every system.
Most companies we meet have three or four of these at once. That’s the jungle. The semantic model untangles it: SAP’s meaning in one tested place, and every spreadsheet, report, tool and AI reads from there — today, and after the move to S/4HANA.
The three answers in the example aren’t bad luck. Reports and AI built straight on SAP tables give likely answers. Wherever the data lands — a data platform, BW, Power BI or a spreadsheet — an extract falls short in three ways. The fix is the same for all three: a semantic model.
Today’s contract price, for example, isn’t stored as one number — SAP derives it each time. An extract can’t run that logic. SAP’s own simulation (via BAPI) can, without creating a document.
Cancellations, movement types, document flow, currencies: correct data still gives wrong answers when nobody has defined what it means. That’s the semantic problem.
Z-tables, custom fields and configuration that departs from SAP standard are where your business really runs. Standard data products and generic models are built for standard SAP — not for yours.
From one warehouse to the next, and on to the customer.
In an extract: separate tables, with movement types nobody has decoded, links nobody has followed — and custom movement types no standard model knows.
With the semantic model: the path modeled once — one traceable flow from receipt to customer, for every warehouse, including the new one.
Contract-based pricing, by customer and product — today.
In an extract: past invoices carry their prices. Today’s contract price doesn’t exist as one number — SAP derives it from condition records, scales and custom routines. The extract has the ingredients, not the recipe.
With the semantic model: today’s prices calculated by SAP’s own pricing simulation — no documents created — and tested against real invoices.
Each of those gaps produces wrong answers. No IT team would accept 98% uptime — the same percentages applied to SAP answers show where the money is, and where the budget for fixing it comes from.
We start from our industry best-practice models for SAP standard — the work, and the value, is in your delta. We build it in the medallion layers your data team already knows — on your data platform, or on N2ONE Core with PostgreSQL if you don’t have one. Silver and Business Vault carry SAP’s rules; Gold serves the answers. The result is your IP, and the foundation that turns AI from a guess into a tool.
SAP tables as they arrive — VBAK, VBRP, KONV, your Z-tables.
Customers, materials and orders joined by business keys, with document flow and full history.
The questions your expert answers in their head — written down and applied the same way everywhere, including your Z-tables and non-standard configuration:
Each metric defined once, with business terms and verified questions — for every report, tool and AI.
Packed tables. On classic ECC databases, pricing conditions (KONV) and accounting lines (BSEG) sit in cluster tables. Generic connectors can’t unpack them.
Fivetran docs →Currency decimals. SAP stores 100 Japanese yen as 1.00. Read without SAP’s currency settings, such amounts are off by a factor of 100.
SAP Community →Leading zeros and blanks. Load document numbers as numbers and joins quietly fail. Let blanks turn into NULLs and filters miss records.
Google Cloud docs →Deletions. Every row needs a change flag. Miss the deletes, and records SAP has removed keep counting in your reports.
Google Cloud docs →Business logic. “Simple replication of tables does not preserve calculations, validations, or conditional rules.”
LeverX →We check the first four in Bronze before anything is built on it. The fifth is where the three gaps begin — and what Silver, Business Vault and Gold are for.
Reports and AI on extracted data~82%of answers right in our model
Reports and AI on the semantic modelMeasuredagainst your books — target three nines
“Verified” means: tested against the reports your finance team already trusts, and signed off by your own data owner — not by a vendor.
See the standard models first. Our SAP Data Model Library: industry best-practice medallion templates with diagrams and semantic definitions — free after a short review.
Explore the libraryEvery step pays off on its own — starting with reporting you can trust. Each one adds to the documented understanding of your SAP data that AI and your S/4HANA migration will need.
07 · WHO WE ARE · N2ONE, THE SAP SEMANTIC EXPERTS
Securing your IP — the meaning of your SAP data, documented and owned by you.
★★★★★ 5.0 Verified client reviews
“…very experienced in this space and is exactly what we needed during our evaluation process.”ERP evaluation consulting
“Very nice, professional, collaborative, and smart.”Enterprise data architecture
Recent work: a federal AI-to-SQL program. The demo ran almost immediately; accurate SQL needed a semantic model.
Your 20 questions show where today’s answers go wrong and what that costs — and they set the scope of the workshop. The workshop solves the immediate need and gives you a data model every next project builds on: new reports, AI, new sites and S/4HANA.
Give us 20 questions your teams ask — ours or yours. We test today’s answers against your trusted figures and show what the errors are worth.
From your team: one sponsor, plus a few hours from the people who own the reports.
Typical questions: revenue by product line vs. the close, customers billed below contract price, stock value by plant, orders at risk this week.
Access: read-only, under NDA.
Five days, one department, scoped by your 20 questions. We build the core semantic model — SAP standard plus your custom Z-tables — and document it in N2ONE Core. It solves today’s need and becomes the foundation for what comes next.
What you get: a working semantic model in N2ONE Core — dynamic, so new reports and AI questions build on it instead of starting over.
Expert time: a few working sessions with your key expert — we do the documenting.
Scope: one department, up to 25 metrics, plus the custom tables they touch.
Our commitment: if the model doesn’t reconcile with your books for the agreed metrics, we keep working until it does — at no extra fee.
Yours to keep: the model and its documentation belong to you — with or without N2ONE Core. Your team can extend it without us.
Where to start: pick the department that matters most now — or the warehouse you’re opening.
Access: read-only on your data platform. In SAP: read-only, plus SAP’s own simulations (such as pricing) that create no documents.
Not ready yet? Explore the free SAP Data Model Library → · Best results when IT and finance join — finance brings the questions, IT the data. Read the white paper →
Short answers to what comes up in the first meeting.
Back to the two steps