smartGRC where your data must stay
We deploy smartGRC in the EU cloud (GCP Frankfurt) - the standard option for most customers. For organisations that cannot send SAP data or AI prompts outside their own network, we also deploy on-premise with a local LLM. Same platform, same AI agents. Your infrastructure - your call.
touch_app Click any element on the diagrams below to see details under the diagram.
Same product. You decide where it runs.
Both options are standard deployments - never a "special project". Same product, same roadmap, same support.
EU cloud (GCP Frankfurt)
smartGRC hosted on Google Cloud, Frankfurt region. Fastest start - 4 weeks to production. Data stays in the EU. GDPR compliant.
- check Standard for most customers
- check AI model managed by us (in the EU)
- check Automatic updates
- check 4-week deployment
On-premise + local LLM
smartGRC in your infrastructure - with no calls out to our cloud. AI runs locally on the model you choose. Standard for banks, public sector, energy majors and large DACH organisations.
- check Data and AI stay in your network
- check Model-agnostic (your LLM choice)
- check Zero outbound calls
- check SAP-partner support standard
The concrete topology (containers / VM / K8s) and system requirements are agreed with your team during scoping - matched to your environment.
Where data and AI traffic flows
Side by side for both options. Click any element to see details under the diagram.
Component and integration topology
Gateway + domain modules as independent containers - identical across both deployment options. Only WHERE the containers and data are hosted changes. Click any component to see its description.
The same set of containers in both options - only the hosting differs:
Google Cloud, Frankfurt
- check Gateway + modules - hosted by us on GCP
- check PostgreSQL + MongoDB - hosted by us on GCP (EU)
- check Keycloak - container managed by us
- check LLM - managed by us, hosted in the EU
Customer infrastructure
- check Gateway + modules - containers / VMs in the customer network
- check PostgreSQL + MongoDB - instances in the customer network
- check Keycloak - can federate local AD/LDAP
- check LLM - local endpoint (OpenAI-compatible), your choice
Who picks on-premise, who picks the cloud
Most customers pick the EU cloud - fastest start, data in Europe, automatic updates. On-premise is for organisations where "sensitive data must remain in our network" is a hard rule:
Banks and financial institutions
SOX 404, MiFID II, GDPR and central bank rules - strict regulator requirements and production separation.
Government and public sector
Data sovereignty requirements, air-gapped landscapes, defense contractors.
Energy and critical infrastructure
NIS2, critical-infrastructure data, utilities and oil & gas.
Healthcare and life sciences
HIPAA, patient data, GxP-regulated pharma manufacturing.
smartGRC in your infrastructure - with no external dependencies
Deployment: smartGRC runs in your environment (containers or virtual infrastructure) with zero calls back to our cloud. You control the data and the update schedule. The specific topology and requirements are agreed during scoping.
Local LLM - model-agnostic: AI agents run on your model of choice via a local OpenAI-compatible endpoint - no prompt, SAP record or decision context ever leaves your network. Run an open-source model (Llama, Mistral, Ollama) or plug in your own private endpoint (e.g. Azure OpenAI Private). Human-in-the-loop, auditable, reversible - same as in SaaS.
Your data, your logs: SAP data, LLM prompt history, model weights and metadata stay on your storage. Nothing is shipped to us. A full audit trail for every AI decision - on your side.
Support and updates: same product, same roadmap, same support as SaaS. The update mechanism is agreed per environment (packages or a secured channel).
Regulated customers trust us
On-premise (and hybrid cloud) is a standard option for our customers - including in heavily regulated industries:
Made in Poland · data stays in the EU · delivered by GRC Advisory (ISO 27001 · SAP Service Partner with 15+ years of experience).
Not sure which option fits?
Book a 30-minute call with our architect. We'll walk through your environment, regulatory constraints and recommend the right deployment.