A/B testing and adaptive experience optimization platform. Processing is limited to pseudonymous experiment data: a randomly generated first-party visitor identifier, variant assignments and goal events, and low-cardinality traits with no direct identifiers. Analysis runs at variant and cohort level only, with aggregate results released under k-anonymity thresholds; the service is not used to profile or target individual users.
Clear communication about data collection, use, and sharing practices.
Collecting only the data necessary for stated purposes.
Demonstrable compliance through certifications, audits, and documented processes.
Technical and organizational measures to protect data integrity and confidentiality.
Ensuring data accuracy, completeness, and currency.
Supporting data subject rights including access, rectification, and deletion.
| Model | Type | Provider | EU AI Act Risk | Open Source |
|---|---|---|---|---|
| Proprietary Reinforcement Learning Models | Custom ML | — | minimal | No |
| Multi-Armed Bandit Engine | Custom ML | — | minimal | No |
| Q-Learning Decision Engine | Custom ML | — | minimal | No |
| Decision Tree Classifier | Custom ML | — | minimal | No |
| Name | Purpose | Location |
|---|---|---|
| Google Cloud Platform | Cloud infrastructure hosting; storage of personal data at rest in encrypted form, including encrypted backups | EU |
| Amazon Web Services | Cloud infrastructure hosting; storage of personal data at rest in encrypted form, including encrypted backups | EU |
SCCs (EU 2021/914, Module Two) with a documented transfer impact assessment; personal data at rest is hosted in EU regions, accessed remotely from the US. Not certified under the EU-US DPF.
Conductrics provides privacy-focused A/B testing and experimentation. Data processing is available in both EU and US regions via Google Cloud Platform and Amazon Web Services. Conductrics applies anonymization and K-Anonymity techniques to protect individual-level data.