Breaking: Competitor had an AI data leak. What should we ask Claw about similar risks?

Just saw the news about the competitor's AI leaking client data. We're evaluating Claw's GRC platform. This makes me nervous.

What specific questions should we throw at their sales team now? Need to cut through the usual "enterprise-grade security" fluff. Looking for concrete, verifiable answers about their AI features. Things like:
- Where our data goes during AI analysis.
- Can we completely disable AI on our instance?
- Who at Claw can access our prompts and outputs.
- Past security audit results related to their AI modules.

Don't want to pay for features that become a liability.