AI, transparency & security
Updated September 13, 2026.
How we use AI
We use AI tools to support research, design, development and content production. People set objectives and review deliverables. Albus videos, 3D portraits and brand visuals include AI-generated or AI-assisted elements. They are not customer testimonials.
What the demonstrations do
Team animations and interactive experiences on this website illustrate possible workflows. They are not a customer portal or an agent team connected to your business. Inputs in these demonstrations are not sent to an AI model.
Access, memory and approvals
For a client deployment, we define approved sources, retained information, authorized people and actions requiring approval. Access should not expand simply because a document or AI response requests it. Integrations and memory features depend on the agreed scope.
Security within the project
The setup plan should address least-privilege access, secrets, malicious or misleading content, appropriate logs, review of important actions and recovery from errors. OWASP generative AI recommendations are useful references, not a certification. Delivered controls and maintenance responsibilities must be agreed and verified for each project.
Privacy and limitations
AI can produce errors, retain incomplete context or misinterpret requests. Consent, minimization, privacy impact assessment and transfer safeguards depend on the project and applicable laws. Confidential information should be provided only after the channel, providers and access rules have been agreed.
Report a concern
You can report a content, privacy or security concern to us. Send a minimal description; do not attach passwords, access keys or customer data. We will agree on an appropriate channel if more details are needed.