AI Transparency Notice
Last updated: Pending publication
PropDev OS includes an AI assistant named Vera that helps you set up projects, answer questions about your appraisal, and surface insights. This notice explains how Vera works and what it does (and does not do) with your data.
Model providers
Vera is built on large language models from Anthropic (Claude family) and may, where explicitly indicated in the UI, route specific tasks to OpenAI. Both providers are listed on the sub-processors page.
What Vera does
- Reads the project context you have open (e.g. budgets, appraisal inputs, milestones) so it can answer questions in context.
- Generates suggestions, summaries, and explanations of metrics visible in the dashboard.
- Stores chat history within your organisation so you can reference past conversations.
What Vera does not do
- Training: Your prompts and project data are not used to train Anthropic’s or OpenAI’s base models. API calls pass the no-training flags supported by each provider.
- Cross-tenant access: Vera only reads data visible to your organisation under our row-level security rules.
- Autonomous action: Vera does not write to your project, send emails, or trigger payments without your explicit confirmation.
Limitations and accuracy
Vera is an AI system and may produce inaccurate, incomplete, or misleading output. Do not rely on Vera’s output as investment advice, legal advice, or as a substitute for professional judgement. Always verify material decisions against your source data and qualified advisers.
Data you share with Vera
Avoid pasting personal data of third parties (clients, contractors, etc.) into Vera prompts unless you have a lawful basis to do so. Vera’s in-app banner reminds you of this on the first message of each session.
Retention
Chat history is retained for 1 year and then automatically deleted by our scheduled retention job. You can request earlier deletion via Settings → Delete Account (deletes everything org-owned) or by contacting privacy@propdev-os.com for targeted chat deletion.
Reporting issues
Hallucinations, biased output, or other AI issues can be reported to privacy@propdev-os.com — we review reports as part of ongoing model monitoring.