Skip to content

vise DMW · Data Migration Workbench

Know what will fail before you load.

Business and IT run the data side of a migration on one screen: extract, validate, profile, map fields and values, simulate the load, post through standard interfaces, report. The whole chain — or the layer in front of the load tool you already have.

Start a 30-day PoC on your data Watch the demo
Since 2017
in production at a global brewing group
8 steps
extract to report
Dozens
of countries migrated to SAP
Silent, 29 s · screens from the Nordic Rollout demo programme · full four-minute demo

01

One workbench. Three ways to bring it into a programme.

DMW sits where data-quality tools and migration tools overlap. Customers use the whole chain, or only the part their integrator is missing — nobody has to change their load tool.

01 Data quality before migration Global brewing group In production since 2017. DMW validates and cleans the master data; the load goes through the existing BODS pipeline. One readiness score instead of twenty trackers. Read more 02 End-to-end migration with AI field mapping Karlsberg Customer and BOM data for the SAP-to-Navision move, delivered by one specialist instead of five. Now the first production user of FieldMap, mapping material data from Business Central to S/4HANA. Read more 03 Standalone data quality Energy infrastructure operator Selected in a public tender ahead of two global data-quality vendors. Rules, scorecards and correction queues — no migration attached. Read more

02 · The eight steps

Every step is governed. Every step is visible.

Pick a step to see what it does and what the screen looks like.

Step 03 · Validate

Rules as code, run on every record, owned by a named person.

Validation rules are small functions with a scope, a severity, a documented reason and an owner. They are versioned, tested in a sandbox and run against the live extract. Business users see the results per object and per field; architects see the source. AI can draft a rule from a sentence — a person activates it.

  • Format, LoV, conditional, custom and aggregate rules
  • Severity, scope and owner on every rule
  • AI-drafted, human-activated
vise DMW validation rules: 156 rules across 119 fields and 7 objects, filtered by site
Validation rules — 156 rules, 119 fields, 7 objects (demo)

03 · Into an existing pipeline

Your integrator keeps the load tool. vise DMW sits in front of it.

Source → DMW (extract, validate, map, simulate) → the load path you already have → S/4HANA. Three ways to wire it, chosen per object. vise DMW keeps the rules, the mappings and the audit trail; SAP keeps the records.

A · DMW loads Built-in loader, one button Posts through BAPI, IDoc or OData/RAP, batch by batch. Every rejected record comes back into the correction queue with SAP’s message; the rest of the batch continues; fixed records are re-posted. One audit trail from validation to load.
B · Migration Cockpit loads Pre-validated files for LTMC / LTMOM DMW delivers the dataset in the Cockpit template format, already mapped to the migration object and already past validation and simulation — so the Cockpit upload is a formality, not the first test.
C · BODS or your own loader Clean dataset, documented rules DMW hands over the validated, mapped dataset as CSV, XML, JSON or staging tables, with the mapping specification and transformation history. This is how our longest-standing customer runs it: DMW as the readiness gate, BODS as the loader.

In every option the load goes through SAP’s standard interfaces, so CVI, customizing checks, BAdI/exits, number ranges and authorizations stay active. DMW never writes to a table directly.

04 · Rules as code

AI drafts the rule. A person switches it on.

Describe the check in a sentence — “a Norwegian vendor’s VAT number must be NO plus nine digits and MVA” — and the DMW assistant drafts the rule, runs it in a sandbox against real records and reports what passed. The rule is a readable function with a scope, a severity and a documented reason. Activating it is a human decision, logged with the user and the time.

  • Every rule has an owner, a version and a sandbox test result.
  • Architects read the source; business owners read the sentence.
  • Nothing the AI writes runs on production data until approved.
DMW assistant: a validation rule drafted from a sentence, tested in a sandbox with passing and failing examples
DMW assistant drafts and tests a rule
Rule card for GL account currency match: scope, documentation and the Python source of the rule
Rule card with scope, documentation and source

05 · Reports

Data quality your business owners can read.

Not a developer’s log. The pages a programme manager, a data steward and a steering committee open every week.

vise DMW summary report: valid records 90.7%, +1.6pp vs last week, biggest improvers and regressions
01 Weekly summary Who improved, what regressed, what is due. The page the programme manager sends on Friday. Valid records 90.7 % · +1.6 pp vs last week · biggest improvers and regressions
vise DMW top offending rules with error counts, six-week delta and trend per rule
02 Top offending rules Rules ranked by errors, with the six-week trend and the owner next to each one. Errors per rule · six-week delta · trend per rule
Apache Superset dashboard on vise DMW data: DQ score, records validated, invalid records, rules with errors
03 Superset dashboards Live BI on the same validated data — KPIs, trends and site comparison without an export. DQ score · records validated · invalid records · rules with errors
vise DMW meeting notes: weekly data quality review with decisions and open points per object
04 Meeting notes and attention points Decisions, open points and owners per project, attached to the data they talk about. Decisions · open points · owners, per object

06 · For SAP architects

Standard interfaces. No surprises in your system.

The questions we get in every architecture review, answered the same way every time.

The full architecture page
01How does DMW load into S/4HANA?
Through standard BAPIs, IDocs and OData only. Never a direct table write. Business Partner goes through CVI like any SAP-approved tool.
02Can it feed Migration Cockpit or BODS instead of loading?
Yes. vise DMW can be the DQ and mapping layer in front of your existing load tool, or the load tool itself. For Migration Cockpit it delivers pre-validated files in the template format; for BODS or a custom loader, the mapped dataset as CSV, XML, JSON or staging tables with the mapping specification. Our longest-standing customer runs it in front of BODS.
03What does Load Sim check?
The simulation runs the dataset against your target’s rules with your customizing: mandatory fields per role and grouping, account groups, tax formats per country, BAdI-level checks — plus the business rules SAP does not know. Because the target system itself validates each record, the simulation returns the same errors the real load would — per record, into the correction queue — without committing anything.
04What do you check that Migration Cockpit does not?
Cockpit, LSMW and BODS validate what SAP knows: formats, mandatory fields, customizing. That answers “will SAP accept it?”. DMW adds the rules SAP does not know — cross-object consistency, country and regulatory checks, finance governance, your own purchasing and sales policies. That answers “will the business work with it?”. Both layers stay on.
05What does the AI touch?
Drafts only: field mappings, transformation rules, validation rules. Every proposal is versioned, tested in a sandbox and approved by a named person before it runs.
06Where does it run?
In the EU, in a dedicated environment per customer (AWS or a dedicated OVHcloud server in Germany) operated by Visehub; connection to your systems through a site-to-site IPsec tunnel or an extract hand-over; named accounts and roles; TLS everywhere, encrypted backups in Frankfurt. ISO/IEC 27001 and ISO 9001. Details on the Trust page.

Works with the rest of the vise family

In the Field Map step · also standalone vise FieldMap AI-drafted field mapping with a confidence score, approved line by line by your consultant. Explore In the Value Map step · also standalone vise LoV-MAP Value harmonisation by business users, reused across sites and waves. Explore After go-live vise Data Run The same validation, simulation and approval for everyday data changes. Explore

See your own data in DMW in 30 days.

One object from one site — customer or customer or vendor master is the usual pick — extracted, profiled, validated, field-mapped, value-mapped and load-simulated against your target. You get the data assessment report, the working project and an effort estimate for the full wave. Your side: an extract or system access, and one data owner for the decisions. Fixed-price consulting; no licence fee for the 30 days. After that: a per-project licence priced by sites and user accounts (the more sites, the lower the price per site) — your IT runs it, we run it under SLA, or hybrid.

Start a 30-day PoC How we work