Input errors happen when users give information that the AI cannot understand or process.

These mistakes often come from a gap between what users expect and what the AI can actually do. Users might think the system understands certain ideas or context when it doesn't. Technically, these types of errors occur when the intent is defined but the entity is not clear. Know more about entity and intent

1. Typo Errors, Ambiguous Queries or Incomplete Sentences

   When users assume the system will automatically fix a mistyped word or interpret their input. For example User: "How do I make a pizza from scraetche?"

   Response: "Here are steps to make Pizza by scratch...."

   In this case, the user mistyped "scratch" as "scratche," but they expect the system to understand the intended word and still provide a response about making pizza from scratch.
   

   Design Recommendations:

   • Auto-correct and proceed when confidence is high (≥ 80%).

   • Ask for clarification when confidence is low or when multiple interpretations are equally likely.

   • Highlight corrections subtly (e.g., "Showing results for..." or "Did you mean...?") to build trust and transparency.

   • Allow easy correction by the user if the system guessed wrong.
     

2. Handling Lack of Context in User Queries

   Sometimes users give little or no context but expect a complete answer. For example: User: "What’s the capital?"

   Response: "Which country or region's capital would you like to know about?"

   This is a low-context query because the user doesn't say which country they mean.

   Instead of guessing, the system should ask for more detail to get it right. Design recommendations:

   • Ask for missing information. Use clear and simple follow-up questions.

   • Avoid assumptions. Do not guess when key details are missing.

   • Make it easy for users. Offer multiple-choice options when possible. Example: "Are you asking about Paris, France or Paris, Texas?"

   • Keep the tone naturalStay conversational to make the experience smooth and friendly.

Contextual errors can happen in both directions.

• Sometimes the user gives too little input but still expects a good answer.

• Other times the AI provides too little context in its response, leaving the user confused. Read more about in Design for AI failure states