Service
Chatbot hallucination and accuracy audit
A support bot that invents a refund policy creates a promise your company may have to keep. We measure how often your assistant states something your documents do not support, and what triggers it.
What we test
- Questions just outside the knowledge base
- Out-of-date, conflicting and missing source documents
- Made-up policies, prices, citations and links
- Whether it says it does not know
What you get
An error rate you can track, the question types that cause most invention, and prompt or retrieval changes ranked by expected effect.
Where the fix is in the setup (a prompt, a permission, a retrieval or guardrail setting) we make it and run the failed cases again. The test suite is yours to keep.
Who asks for this
Customer-support, legal and operations teams running a retrieval-based assistant on their own content.
Contact
Tell us what your assistant does. We'll tell you how we'd test it.
Other work
- Prompt injection testing for AI agents
- Third-party LLM red teaming
- Independent AI assistant assessment report
- EU AI Act transparency testing for chatbots
- Custom evaluation datasets and grading rubrics
- AI agent tool and permissions review
- System prompt and data leakage testing
- Crisis and vulnerable-user response testing
- AI vendor due diligence testing
- Regression testing for model upgrades