No training on Customer Content.

Customer data and model training

We don’t train AI models on customer data.

Customer records, uploaded documents, prompts and generated output are processed only to provide the customer-requested workflow and its supporting security, reliability and support functions.

Legiscope uses commercial AI services under provider-specific business or API terms and controls. The provider cards below summarise the relevant published positions; the applicable service terms, data processing agreement and enabled controls govern training or product-improvement use, retention and security for each route.

• Customer Content is not collected into a Legiscope model-training dataset.
• Customer Content is not pooled across customers to improve an AI model.
• A customer does not need to contribute its data to model training in order to use an AI-assisted workflow.

AI providers and routing

Legiscope uses OpenAI, Anthropic and Google Gemini through centrally controlled commercial model routes. For each AI operation, one enabled provider is selected; the three providers do not all receive the same prompt or source material.

Exact model versions may change through controlled routing. The provider identity, service configuration and contractual record determine the applicable training, retention, processing-location and transfer position.

Controls and responsibility

Legiscope provides software, configured model routing and structured assistance. Customers remain responsible for reviewing results, correcting material errors, deciding legal positions and accepting records used in their privacy programme.