Contradiction Handling

When you share information that conflicts with existing knowledge, your AI employee detects the contradiction and alerts you. For example, if you previously said your standard rate is $150/hour and now say it is $175/hour, the AI will flag this and ask which value is correct. It does not silently overwrite existing knowledge — transparency is built into the system. This prevents accidental misinformation from corrupting your knowledge base.

When a contradiction is detected, you have several options. You can confirm the update, which replaces the old information with the new. You can keep both entries with different contexts — for example, different rates for different client tiers. You can dismiss the new information if it was stated in error. Or you can ask the AI to show you all related entries so you can make an informed decision. The AI records your resolution decision for future reference.

The knowledge base maintains a version history for updated entries. When information changes, the previous version is archived rather than deleted. This means you can always trace back to see what information was current at any point in time. If you quoted a price to a client three months ago and the price has since changed, you can verify what the price was at that time. This audit trail is especially valuable for pricing, contractual terms, and business decisions.

To minimize contradictions, try to be consistent in how you communicate information to your AI employee. When making changes, explicitly state that you are updating existing information — "update our rate to $175" is clearer than just stating "$175." Periodically review your knowledge base by asking the AI to summarize what it knows about key topics. This helps you catch any inaccuracies early and keeps your business information reliable and current.