Executive Intelligence — one cockpit across SAP FI, CO, MM, SD, PP, HR

What's in today

• Reads SAP FI, CO, MM, SD, PP, HR via whitelisted Z-OData services
• Anomaly engine with severity scoring
• AI narratives per module
• Natural-language rule creator (backend live)

The anomaly engine across modules

The same numeric-guardrail pattern runs across six SAP modules.

• FI — cash and GL anomalies
• CO — cost-center deviations
• MM — purchasing and stock signals
• SD — order and billing patterns
• PP — production variance flags
• HR-finance — headcount and payroll deltas

Each anomaly fires with severity (critical, high, medium, low). Each comes with a “why” panel — what fired, what the threshold was, what the underlying SAP figure is. CFOs see the story; auditors verify the numbers.

Natural-language rule creator

Type a rule description in plain English; the backend creates the rule. “Alert me when AP aging over 90 days exceeds 5% of total AP” becomes a rule that runs against your Z-OData reads. Rules are previewed against historical scans before going live.

Frontend rule creator UI is rolling out; backend is live today. Ask us about pilot access if you want to script rules now.

AI narratives, validated against your numbers

Every figure in a module narrative is checked against the underlying aggregate. >5% drift rejects the narrative — it doesn't ship to the dashboard. The CFO reads a story whose numbers always tie to the dashboard charts. The auditor opens the same period and finds the figures match the underlying SAP data.

What's coming

Coming soon: executive chat agent, narrative feedback collection, expanded SAP coverage. Ask us about pilot access.

Frequently asked questions

Which SAP modules are read?

FI, CO, MM, SD, PP, and HR — through whitelisted Z-OData services your Basis team controls.

How fresh is the data?

Cockpit reads run on schedule against the OData services. Frequency is configurable per tenant; freshness depends on the source schedule.

Can I create custom rules?

Yes. Natural-language rule creator backend is live; UI is rolling out.

Are the AI narratives ever wrong on numbers?

They're validated against the source aggregate before display; >5% drift gets the narrative rejected. So no — what you see has been numerically reconciled.