This document is pending formal legal review and may be updated before general launch.
TenderLock uses AI to help people run better tenders. This policy explains what the AI does, what it does not do, and how we keep it accountable.
1. AI assists. People decide.
AI drafts tenders, reviews them, checks submissions and suggests scores. It never publishes a tender, excludes a supplier, records a final score or awards a contract. Those actions are always taken by a named user and recorded.
2. Evidence, not assertion
AI findings link to the source they rely on, such as a requirement, an answer or a page in a document. Where evidence is missing or a document cannot be read, the AI says so rather than guessing.
3. Published criteria only
Submissions are assessed only against the requirements and weighted criteria published to suppliers.
4. Provenance
For each AI action we record the function used, the model, the sources provided, a fingerprint of the output and what a person did with it. Buyers can export this record.
5. Data use
Your content is processed only to carry out the action requested. It is not used to train AI models. Suppliers' content is never used to assist competing suppliers.
6. Providers
AI features currently use OpenAI models under business terms that exclude training on your data. We may change provider, and will keep this policy up to date.
7. Your responsibility
Review AI output before relying on it. You remain responsible for your procurement decisions and for compliance with any rules that apply to you.