What AI-powered features are available to users?
Feature/ Description:
AI Assistant
An intelligent chatbot that answers questions about your diagrams and makes edits based on your instructions.
AI Imports (PDF/Word/Image/Text/CSV/Excel)
Automatically analyses uploaded files and converts them into a visual entity-relationship diagram.
How AI Modes work
Your System Admin controls how your data is processed through two modes:
Your Region
Data is processed in data centres close to your agreed hosting region. Currently supported by Azure OpenAI (GPT models) only.
Global
Data may be processed across global infrastructure to leverage the latest AI models, including Claude (Anthropic) and Gemini (Google).
AI providers
StructureFlow has integrated with three leading AI providers. All are bound by formal data processing agreements.
Microsoft Azure OpenAI
GPT models, processed within the Azure environment.
Claude in Microsoft Foundry
Claude models, processed on Anthropic's infrastructure.
Google Vertex AI
Gemini models, processed on Google's infrastructure.
Data privacy & security
No data retention for AI training
Your data is never used to train or improve AI models.
Prompt logging
StructureFlow does not log any user prompts or inputs.
Storage
StructureFlow does not store any user prompts or inputs.
International transfer protections
Where data is processed outside the EEA, transfers are governed by EU Standard Contractual Clauses (SCCs) under each provider's Data Processing Addendum.
Disabled by default
AI features are turned off unless explicitly enabled by your administrator.
Controls for your organisation
AI is disabled by default across all user accounts.
System Administrators can enable AI and select the appropriate mode via the admin settings panel.
All users must log out and back in for changes to take effect.
Self-hosting customers can provision their own AI models independently.
EU AI Act Compliance
StructureFlow operates as a Deployer under the Act, using AI systems provided by Microsoft, Anthropic and Google.
Our AI system is classified as minimal risk - it supports diagramming tasks and does not impact individuals' rights, health or safety.
We uphold voluntary commitments to transparency, responsible usage, ongoing monitoring and human oversight of AI outputs.