01Watch every KPI. The cash position agent sees the 13-week forecast fall below its policy floor in week 6.
Cash position agent · AI BI & DW
Procurement, finance, people, assets, planning and sales on one data layer. Agents act on your KPIs, and a person approves every move.
They share one identity layer, one event bus and one AI-native data warehouse, so a customer in CRM is the same customer in ERP, in EPM and in an agent's memory.
Sourcing to settlement, with supplier intelligence built in.
The financial and operational backbone, with agents built in.
One HCM. Country packs, not custom code.
Asset health, work execution and spare parts in one model.
Plan, consolidate, forecast, narrate.
Sales, marketing and service on one customer record.
One content service for every other module.
One semantic layer. Every metric, exactly once.
Freight rose 14% on two routes and hardware discounts took 0.6 points.
The CEO, CFO, COO and CIO work from one platform and one set of numbers. When an agent proposes a move, the person who owns the KPI approves it.
Every move an agent makes is a real API call against the same services your people use, under the same permissions and the same audit log. Shown here for Vellmar Group, an invented company.
01Watch every KPI. The cash position agent sees the 13-week forecast fall below its policy floor in week 6.
Cash position agent · AI BI & DW
02Find the cause. 73% of the gap traces to three overdue customers. A collection sequence is drafted for each.
DSO and collections agents · CRM · ERP
03Plan the response. Supplier deferrals and stock markdowns are scored and drafted within your governance limits.
Working capital agent · Procurement · Treasury
04One approval. Every plan arrives as one approval, with lineage and segregation-of-duties checks already run.
CFO · Person in the loop · BPM
05Execute and learn. Collections go out, payables reschedule, markdowns post. Every step is logged and the KPI returns to range.
Orchestrator · Audit log
REST, GraphQL and gRPC on every module. Build your own portal, embed our components or plug into Teams. Your interface, your brand and your roadmap, on our backend.
OpenAPI 3.1 · BPMN 2.0 · REST · GraphQL · gRPC
01
Your portal posts a journal entry. The call is the same one your engineers read in the docs.
POST /v1/erp/journals
{ "entity": "Vellmar d.o.o.",
"amount": 48200.00, "currency": "EUR" }02
Every module is described in OpenAPI 3.1, so the call is checked before anything moves.
JournalEntry · OpenAPI 3.1
03
The caller gets exactly the permissions of a person in that role, with segregation of duties checked.
Role AP clerk · Vellmar d.o.o.
04
One event tells every module that cares. Planning and BI update without a nightly batch.
journal.posted
Delivered to
05
The ledger posts the entry in the same model your people work in.
06
Every step lands in one audit log, whoever made the call.
09:41:07.214 portal:vellmar-finance
journal.posted JE-2026-04417
trace 7f3c…a91e
From your portal into DivetIQ and back. Every step is the same for a person, a portal or an agent.
Audited for security, built for data protection, ready for EU sovereign cloud, and described in open standards your architects already read.
Your BI tools and data platforms read the same metrics from one open lakehouse, and DivetIQ runs on any cloud, on Kubernetes or on premise.
Pricing model
Pay per use: metered transactions, API calls and agent actions
Architecture
Headless and API-first: REST, GraphQL and gRPC
Intelligence
Specialist agents that execute, with an audit trace
Module breadth
Eight modules under one semantic layer
Time to value
Rollout in 1 to 3 months
Data sovereignty
EU sovereign cloud (GAIA-X aligned), air-gap option
Lock-in
Open table formats (Iceberg, Delta)
Based on public pricing and product documentation.
We meter transactions, API calls, agent actions and storage. If a module is not used this month, it is not on this month's invoice.
Per transaction, per API call, per agent action, per gigabyte. Only chargeable use appears on the bill.
Modules turn on and off independently. Pause CRM in the off-season, scale Procurement for a sourcing window.
Real-time cost per module, team, workflow and query, with alerts for unusual spend.
Switch a module off to see the month's invoice without it.
Sample monthly invoice
Worked example for a 250-employee company, March 2026
INV-2026-03-247
Vellmar Group
| HCM247 active employees, 1 payroll run | 247 active employees, 1 payroll run | €1,234.50 | |
|---|---|---|---|
| Procurement4,800 invoices matched, 12 sourcing events | 4,800 invoices matched, 12 sourcing events | €890.00 | |
| ERP92,400 journal entries, 3,200 sales orders | 92,400 journal entries, 3,200 sales orders | €1,470.30 | |
| BI / DW1.2 TB scanned, 18,400 queries | 1.2 TB scanned, 18,400 queries | €640.20 | |
| CRM12,800 active contacts, 38,000 emails | 12,800 active contacts, 38,000 emails | €512.40 | |
| EAM1,140 active assets, 720 work orders | 1,140 active assets, 720 work orders | €388.10 | |
| ECM740 GB hot, 4.2 TB cold, 3,200 OCR pages | 740 GB hot, 4.2 TB cold, 3,200 OCR pages | €212.80 | |
| Agents1,840 agent actions, 90 approvals | 1,840 agent actions, 90 approvals | €296.50 |
Every module exposes its capabilities through APIs and events, with no hardcoded interface. You use DivetIQ through any front end: web, mobile, partner portals, embedded components, voice or AI agents.
Specialist agents evaluate KPIs against thresholds, trace variance across modules, draft the remediation and execute it through the same APIs your people use, within governance limits you set per agent and per role.
Transactions, API calls, agent actions, storage tiers, analytical scans and integration runs are metered. Modules turn on and off independently, and real-time cost observability is included.
Any of them. All eight modules, from Procurement to the AI BI and Data Warehouse, deploy independently.
Yes. Public cloud SaaS, single-tenant SaaS, private cloud on AWS, Azure or GCP, on-premise Kubernetes or OpenShift, hybrid, EU sovereign cloud (GAIA-X aligned) and air-gapped private cloud are all supported.
Power BI, Tableau, Looker, Excel and Google Sheets, plus any JDBC or ODBC client. The lakehouse is also readable by Spark, Trino, DuckDB, Snowflake, BigQuery and Databricks under governed sharing.
Mid-sized companies: roughly 100 to 2,500 employees, EUR 20 to 500 million in annual revenue and up to 25 legal entities. It leaves out Fortune 500 complexity and adds the mid-market capabilities enterprise vendors charge for but rarely deliver well at this scale.