emeryAI speaks on your behalf to your customers. That’s why we protect both: your data at every level of the architecture and your brand in every single response.
emeryAI processes your data within the EU. Optional features such as web search can be used by providers outside the EU; this is decided per Workspace. For your website visitors: store only as much as necessary, and always transparently let them speak with an AI.
Hosting and processing in the EU
Data in Frankfurt (eu-central-1), including backups. Speech processing via Mistral (France) and Anthropic via AWS Bedrock in the EU region.
No training with your data
Your data will not be used to train large language models (LLM).
Total control
Granular releases per knowledge document, all data can be viewed in the dashboard. Delete and export at any time.
Data about your website visitors
Only what is necessary for response and evaluation is stored. Visitors always see that they are talking to an AI.
Security?
Not a feature. Architecture.
Security
Invisible in everyday life. Verifiable in the audit.
From encryption to access control, it relies emeryAI on established security standards. Each workspace is isolated at the database level, access is governed by role-based permissions, and system events are monitored and logged in real time. This keeps your data protected, separate, and traceable.
Encryption
All connections to emeryAI run are encrypted using TLS 1.2 or higher. The database, file storage, and backups are encrypted using AES-256 and located in Frankfurt.
Access control
Only a small group within emeryAI the team has access to your Workspace; this is exclusively for support and operations purposes. According to our internal guidelines, we only access your content after you have given us permission to do so.
Separation of clients
Isolation per Workspace at the database level via Row-Level Security (RLS). No access to other workspaces.
Monitoring and Abuse Detection
Automated protection mechanisms detect patterns of abuse and intervene early.
Fire safety
Your brand. In every answer.
emeryAI speaks in your name, therefore it speaks only about what you release, and only in the way your brand speaks. Answers are based on your own sources of knowledge; topics outside the established framework are diverted, and wording, tonality, and no-gos are firmly anchored in the configuration.
Your expertise. Made useful.
Answers follow a knowledge hierarchy that you can access internally. Each statement can be traced back to its source and is presented in a transparent manner.
Topic control
You determine the topics you will emeryAI discuss. Requests outside the framework are recognized and redirected.
Protection against manipulation
Multi-stage protection mechanisms against prompt injection, role switching, and social engineering attempts that are continuously being developed.
Brand loyalty
We determine the wording, tonality, and no-gos of your brand together with you. These are considered fixed guidelines for every response.
Governance
Made in Germany. Ready for the AI Act.
From the contract to the regulation, the requirements applicable to AI in the European customer dialogue have been taken emeryAI into account. You receive an AVV in accordance with GDPR, including a list of Toms and sub-contractors, an assessment according to the EU AI Act, and operation on certified infrastructure, developed in accordance with German and European law.
Order processing
AVV in accordance with Article 28 GDPR including TOMs and an overview of all subcontractors.
EU AI Act
Classification according to the EU AI Act and implementation of the transparency obligations towards website visitors.
Certified infrastructure
Hosted at Supabase, AWS, Vercel, and Cloudflare, all certified according to ISO 27001 and SOC 2 Type 2.
Developed in Germany
Development and operation in accordance with German and European law.
Your expertise is your advantage. We ensure that this remains the case.
Consent is required to stream this video via service Bunny Stream. It may involve sharing data with third parties.