Question 1

What is Virtual Data Platform?

Accepted Answer

Virtual Data Platform gives every user — and every AI assistant — live, governed access to enterprise data in the tools they already use, without copying it. Instead of moving data into a new store, the platform brings the query to the data: it reads live from your source systems and returns only the results. Connect everything, copy nothing.

Question 2

What is data virtualization, and how is it different from a data warehouse or data lake?

Accepted Answer

Data virtualization is the underlying technology: rather than physically replicating data into a central store, it leaves data where it lives and queries it in place, on demand. A warehouse or lake copies data first and queries it later, so answers are only as fresh as the last load — the platform queries the source live, so the answer reflects the current state. It also spares you building and maintaining a pipeline for every new question.

Question 3

Does it replace our data warehouse or ETL pipelines?

Accepted Answer

No — and it isn't meant to. Curated history, heavy batch transformation and regulated retention still belong in a warehouse or lake. Virtual Data Platform fits over your existing landscape and spares you the next copy: your warehouse, lake, SAP systems, databases and files simply become sources it reads live, alongside everything else.

Question 4

Does Virtual Data Platform copy our data?

Accepted Answer

No. Data stays in your source systems. Queries execute live at the source and only the relevant data is retrieved — nothing is persisted on the platform.

Question 5

How is access governed?

Accepted Answer

Every query runs under the signed-in user's own identity — never a shared service account. The user authenticates with Microsoft Entra ID, and an Agent translates that identity into whatever each source expects (for example an SAP Logon Ticket, SAML or existing credentials). Your source systems' own permissions then do the filtering, so there is no second authorization layer to maintain.

Question 6

Is anything stored or persisted on the platform?

Accepted Answer

No. Results exist only in memory, inside the requesting user's ephemeral, isolated sandbox, and are destroyed after use — nothing is written to disk on the platform. The platform holds the definitions of your Virtual Datasets, not your data.

Question 7

Is our data encrypted — and can the platform or its infrastructure see it?

Accepted Answer

Yes. Every connection is encrypted with TLS, and for end-to-end-capable clients (the Portal, the Excel add-in, R and Python) the payload is encrypted from the Agent all the way into your client. The decryption key is generated in your client and never leaves it, so everything in between — including our routing and relay infrastructure — only ever sees ciphertext, never your data.

Question 8

How are administrators and privileged roles handled?

Accepted Answer

The two privileged roles — VDP_ADMIN (manages Locations, Agents, Systems and licensing) and VDP_ENTERPRISE_CREATOR (curates the company-wide workspace) — are assigned in your own Microsoft Entra ID, via the Virtual Data Platform enterprise application. That means they are governed where your identities already live, so your joiner/mover/leaver processes apply automatically. Everyday User and Creator licenses are managed by your admin in the Portal.

Question 9

How do workspaces keep access under control?

Accepted Answer

Virtual Datasets live in workspaces: a private workspace is each user's own playground, shared workspaces are managed by teams with read access gated through your existing Entra ID groups, and the global MyCompany workspace is curated and visible to every licensed user. Access is controlled through the groups you already manage — there is no parallel permission system.

Question 10

How do you reach data in other networks or behind firewalls?

Accepted Answer

A lightweight Virtual Data Platform Agent sits inside each network and connects outbound to the platform, waiting for work — the platform never reaches into your network. Connections are certificate-based and encrypted. This lets the platform query data across separate networks that have never been (and never will be) bridged, while keeping each one isolated.

Question 11

Do we need to open inbound firewall ports?

Accepted Answer

No. Because Agents connect outbound only, there are no inbound firewall rules to add and no ports to expose. Isolated networks stay isolated; the data still flows, governed.

Question 12

Which source systems can you connect to?

Accepted Answer

Agents run Handlers for SAP (BW, HANA, Business One), databases and warehouses (Microsoft SQL Server, Snowflake, ODBC), business applications (Salesforce, Workday, Azure DevOps), the Microsoft 365 and BI stack (SharePoint, Microsoft Graph, Power BI semantic models) and open interfaces (OData, web sources). Anything reachable over ODBC or OData can be connected, and new Handlers ship regularly.

Question 13

How does cross-company or cross-network access work?

Accepted Answer

Place an Agent at each location — headquarters, a subsidiary, a plant, even a partner's network — and the platform can combine their data into one consolidated view. Every party queries under their own identity and sees only the datasets they are allowed to see, so collaboration happens without bridging networks or handing over copies.

Question 14

Which tools can consume the data?

Accepted Answer

The ones your teams already use. There is an Excel add-in (formulas and a point-and-click wizard), native R and Python packages, and — through the Universal Connector's virtual SQL Server endpoint — any tool with a Microsoft SQL Server driver, including Power BI, DBeaver and Tableau. A REST/OData endpoint and an MCP server cover everything else, including AI assistants.

Question 15

How do AI assistants like Microsoft Copilot access our data — and is it governed?

Accepted Answer

The platform exposes your Virtual Datasets to MCP-capable assistants (such as Microsoft Copilot or Claude) as tools. The assistant inherits the signed-in user's identity, the same auth translation applies, and every query runs live through that user's ephemeral sandbox — so an assistant can only reach the data that user is allowed to see, and every query is attributable.

Question 16

Is our data used to train AI, or stored by the assistant?

Accepted Answer

No. Queries run live and the results pass through the user's ephemeral sandbox without being persisted — there is no copy for a model to train on. Assistants can only see the Virtual Datasets you license and expose.

Question 17

Can business users build datasets without coding?

Accepted Answer

Yes. The low-/no-code Modeler lets business users build and combine Virtual Datasets — parameterized Functions, composed Models and reusable Value Lists — without programming, all under their own permissions. Parameters appear as form fields with dropdowns in Excel, so people can answer their own questions without an IT ticket. Teams that prefer code can drop in R or Python Code Files.

Question 18

Where is Virtual Data Platform hosted?

Accepted Answer

On Microsoft Azure, in the EU. Each customer gets its own tenant endpoint (for example yourtenant.eu01.virtualdataplatform.com). For the strictest environments, Virtual Data Platform Dedicated runs on a dedicated backend cluster with Azure Private Endpoints, so traffic never traverses the public internet.

Question 19

Is it GDPR-compliant, and where does our data reside?

Accepted Answer

The platform is built for GDPR and hosted in the EU on Microsoft Azure, with Microsoft acting as processor under its Data Protection Addendum. And because queries run live with nothing copied, your data stays in your own source systems — there is no new store of record to secure or to keep in region.

Question 20

How is it licensed?

Accepted Answer

Two license types: a VDP_User license to view and run Virtual Datasets, and a VDP_Creator add-on to build and model them (every Creator also holds a User license). Your admin assigns them in the Portal's license management in a couple of clicks; the privileged VDP_ADMIN and VDP_ENTERPRISE_CREATOR roles are assigned in your Entra ID instead.

Question 21

How do we trial it or see a demo?

Accepted Answer

Book a live demo and we will connect one of your systems and query it together in a 30-minute session — in Excel, Power BI, R or through an AI assistant. After the demo, trials run in your own isolated tenant, which we set up with you on a short call.

Frequently asked questions.

Concept & positioning

Security & governance

Connectivity & architecture

Tools, AI & self-service

Hosting, compliance & licensing

See your own data live in 30 minutes.