Linos Cleanse.
Cleansing the source data before it hits S/4HANA.
Brownfield migrations stall on data quality. Linos Cleanse profiles your source ECC master data, finds the issues that will block or corrupt the migration, and proposes specific fixes — with Claude doing the pattern-matching that Excel macros can’t. Compresses the 4-6 months of pre-DMO cleansing into 6-8 weeks.
Six problems Linos Migrate detects but doesn’t solve.
Migrate is the cockpit. It tells you the program is blocked because CVI sync is at 62%. Cleanse is what brings CVI sync to 99% — without four months of spreadsheet work.
Pre-DMO cleansing eats 4-6 months
Teams hand-clean Excel exports because there's no tool that profiles, dedupes, and standardizes SAP master data at scale with AI.
Duplicates that regex can't catch
"International Business Machines" vs "I.B.M." vs "IBM Corp" — same entity to a human, invisible to a fuzzy-string library.
CVI gate stuck below 99%
BP grouping issues, number range overlaps, missing required fields. Linos Migrate flags it; Cleanse fixes it.
Reference data mapping is guesswork
Legacy tax codes → new codes. Account groups, profit centers, plants. Hundreds of mappings, hand-built in Excel, no audit trail.
Aged open items piled up
Years of AR/AP that should have been cleared, written off, or reclassified. Migrating them as-is corrupts the new system's books.
No audit trail
Compliance asks who changed what, when, why. Excel can't answer. Cleanse logs every applied change with actor, timestamp, before/after.
Eight modules. AI in every one.
Each module owns one cleansing workstream end-to-end and feeds the workbench, where humans review and approve.
Master Data Profiler
Reads KNA1, LFA1, MARA, BUT000, CSKS, etc. Surfaces issues per table on four dimensions: completeness, validity, consistency, conformity. Generates a quality score per object.
Duplicate Detective
Two-stage: blocking + Claude pairwise scoring. Catches duplicates that look identical to humans but differ in punctuation, spacing, or transliteration. AI explains why.
Standardizer
Address normalization · phone E.164 · tax ID format · name canonicalization. Built-in rules + AI-generated rules tailored to your data.
Completeness Auditor
Required fields missing, mandatory relationships broken, dependent fields inconsistent (e.g., LAND1=US but REGIO=Berlin).
Open Items Triage
Aged AR/AP, parked documents, blocked invoices. Claude classifies each as CLEAR / WRITE_OFF / RECLASSIFY / ESCALATE / IGNORE with reasoning attached.
Reference Data Mapper
Legacy codes → S/4-target codes. Tax codes, account groups, profit centers, plant codes. AI proposes mappings with confidence scoring; humans approve.
AI Rules Generator
Provide 10+ examples of dirty/clean records. Claude analyzes the patterns and emits a deterministic rule (JS or SQL) plus estimated coverage and edge cases.
Cleansing Workbench
Reviewer's UI for all proposed changes. Side-by-side BEFORE/AFTER, batch approval by rule or confidence threshold, audit-logged application via RFC/OData/LSMW.
The two products work as one.
Linos Migrate flags the data quality blockers. Linos Cleanse fixes them. Same connectors, same auth, same audit log. Migrate’s Pre-Check and CVI tabs link directly into Cleanse’s workbench.
Mock · RFC · OData v4
Same SAP connector layer as Linos Migrate. Configure once, both products use it.
SSO across both
Single sign-on via your existing IdP (Okta / Azure AD / SAP IAS / XSUAA on BTP).
One audit trail
Migrate writes phase transitions; Cleanse writes cleansing actions. Both go to the same audit log — exportable to your SIEM.
Sold as an add-on to Linos Migrate.
Cleanse is licensed alongside Migrate. Most customers buy them together because the value compounds — Migrate without Cleanse hits the data wall in week 3.
Migrate alone
The orchestration cockpit. You handle the cleansing yourself, with your existing tools.
Migrate + Cleanse
Both products together. Linos Cleanse profiles, dedupes, and standardizes the source data driving Migrate's gate compliance.
Cleanse standalone
For customers who want to cleanse master data without a full migration program. Same product, smaller scope.
See Cleanse in action.
The demo loads with 5,000 dirty customers, 2,500 dirty vendors, and 800 materials — realistic distributions of duplicates, malformed addresses, missing fields, and aged open items. Click around: profile, detect duplicates, generate AI rules, approve fixes in the workbench.