AI Terms
Last updated: May 2026
1. AI-Assisted Workflows
Foixar uses AI-assisted workflows to help generate specifications, code scaffolds, governance reports, decision summaries, audit context, and related engineering artifacts. AI outputs are intended to support engineering judgement, not replace it.
2. Human Review
Customers are responsible for reviewing, testing, approving, rejecting, and deploying AI-generated outputs. Foixar agent findings are advisory unless a customer has configured a workflow or external system to treat them as blocking.
3. Bring-Your-Own Model Providers
Foixar supports tenant-selected AI model providers. Customers are responsible for the provider terms, billing, quotas, region selection, retention settings, model behaviour, and compliance obligations associated with those providers. Foixar may send prompts, context, and outputs to the tenant-selected provider as needed to provide the Service.
4. Foixar IP and Customer Data
Foixar owns its proprietary agent logic, model instructions, validation methods, orchestration methods, rule schemas, audit infrastructure, and governance engine components. Customers retain ownership of Customer Data and outputs derived from Customer Data, subject to Foixar's rights in the platform and underlying technology.
5. No Model Training Without Consent
Foixar does not use Customer Data to train AI models without explicit customer consent. Tenant-selected AI provider processing is governed by the customer's relationship with that provider and the customer's provider configuration.
6. Accuracy and Risk
AI outputs may be incomplete, incorrect, insecure, or unsuitable for a particular environment. Customers should validate generated code, specifications, reports, and governance findings before relying on them for business, engineering, compliance, or security decisions.