This pattern involves clearly conveying what an AI model can and cannot do, including its knowledge boundaries, capabilities, and limitations. For example, Claude AI states its knowledge cutoff when the question falls outside its knowledge domain

Why use it Transparent communication builds user trust, sets appropriate expectations, prevents misuse, and reduces frustration when the model fails or behaves unexpectedly.

How to use it:

• Explicitly state model limitations. Show contextual cues for outdated knowledge, lack of real-time data, inability to provide medical or legal advice.

• Provide fallback or escalation options when the model cannot help. For example, an AI finance system can say “I’m not trained to give stock trading advice, but I can help you explore mutual fund options or explain how they work. Want to start there?”

• Make limitations visible in product marketing, onboarding, tooltips, or response disclaimers.