Connects to the systems you already run.
ALTeX Hub reads from 75+ enterprise systems, read-only by default. Nothing to rip out, nothing to re-platform.
01Systems we connect to
| Category | Systems |
|---|---|
| ERP | SAP (ECC and S/4HANA) · Oracle (E-Business Suite and Fusion) · Microsoft Dynamics 365 · JD Edwards · NetSuite · Tally |
| CRM and sales | Salesforce · HubSpot · Zoho CRM |
| HR and payroll | Workday · SAP SuccessFactors · Darwinbox |
| Travel and expense | SAP Concur |
| Procurement | SAP Ariba · Coupa · Jaggaer |
| Data platforms | Snowflake · Databricks · SQL databases · data lakes and warehouses |
| Banks and markets | Bank statements and feeds · market data · registries, tax portals and sanctions lists via APIs |
| Documents and email | PDF, Word and scanned documents · email archives (for MailVerse and LawRence) |
Most of the evidence sits below the waterline
Dark data is information a business collects and stores while it operates, then never uses again. Fraud leaves its footprints there.
| Source | Example | What it can reveal |
|---|---|---|
| SAP change and security logs | Change documents, the security audit log, firefighter sessions | Bank details changed before a payment; overrides by privileged users |
| Contracts and price agreements | PDF contracts, escalation clauses, rebate tiers | Billing above contract; rebates paid on volumes that do not qualify |
| Scanned invoices and receipts | Images, OCR text, metadata | Duplicates with altered numbers; forged or reused documents |
| Email and workflow approvals | Approval timestamps, delegations, comments | Approvals after the event; one approver for everything |
| Plant, quality and IoT data | MES output, scrap weights, machine hours | Scrap theft; output that does not match material consumed |
| Warranty and dealer portals | Claims, serial numbers, service reports | Repeat serial numbers; claims after expiry; partner outliers |
| Spreadsheets outside the ERP | Provision models, manual calculations | Management override; numbers no system can trace |
| Supplier and third-party data | Registries, sanctions lists, ESG and sub-tier data | Shell suppliers; links to employees; supply-chain risk |
About 55% of an organisation’s data is dark, most of it log data (Splunk). Around 80% of new enterprise data is unstructured: documents, images, email and logs.
Four ways to connect
Read-only database or API
The standard route for live monitoring.
Scheduled extracts
Secure file transfer on a schedule you set.
JDBC and BI connectors
For data lakes, warehouses and BI tools.
Secure upload
For look-backs and pilots, before any live connection.
The SAP data the tests need, read-only
| Area | SAP sources (examples) | What they hold |
|---|---|---|
| Vendors | LFA1, LFB1, LFBK, ADRC; BUT000 in S/4HANA | Vendor master, bank details, addresses |
| Purchasing | EKKO, EKPO, EKBE | Orders and their history |
| Invoices and payments | RBKP, RSEG, BSEG; ACDOCA in S/4HANA | Invoices, postings and payments |
| Change logs | CDHDR, CDPOS | Who changed what, and when |
| Inventory | MSEG, MARD; MATDOC in S/4HANA | Movements and stock levels |
| Sales and credit | VBRK, VBRP; KONV or PRCD_ELEMENTS | Billing, credit notes, pricing conditions |
| Access | USR02, AGR_USERS, AGR_1251 | Users, roles and authorisations |
| HR, pseudonymised | PA0001, PA0006, PA0009 | Assignment, address, bank details |
Examples for SAP ECC and S/4HANA. The exact sources are agreed for each entity.
05From access to first findings
- 01Scope
Agree the systems, entities and period
- 02Connect
Read-only access or extracts
- 03Profile
Check completeness and quality
- 04Map
Fit your data to the insight library
- 05Validate
Reconcile to your ledgers, then go live
Why dark data stays dark, and how to switch on the lights
Fragmented
Many ERPs, plants and countries, and no single view of a vendor or a part.
- Response: one common data model, read-only connectors, and one region first.
Unstructured and multilingual
PDFs, scans and email in many languages.
- Response: LLMs read documents with confidence scores; people check the uncertain fields.
Poor quality
Duplicate vendors, missing fields, free-text descriptions.
- Response: treat data-quality tests as findings and fix the master data at source.
Private by law
GDPR, works council rights, India’s DPDP Act.
- Response: pseudonymise, use only the fields you need, and agree the scope before the first run.
Untraceable
No one can prove where a number came from.
- Response: trace every row to source, hash the inputs and make every run replayable.
The 80% data trap
Pilots stall because each audit rebuilds its own data.
- Response: reusable pipelines and a test library, not one-off scripts.
07FAQ
Do you write back to our ERP?
No. Connections are read-only. Actions such as holding a payment happen in your systems, by your people.
Our system isn’t listed. Can you still connect?
If it has a database, an API or an export, we can usually connect. Tell us what you run.
How long does connection take?
Days for standard ERPs; longer for heavily customised systems.
Do you need our data team’s time?
A few hours to grant access and answer questions. We do the mapping and validation.
Will it work with your systems?
List your systems and we’ll confirm within two business days.