Can you give an example of a question that you would google, but your colleagues would "Ask Claude" and they get a hallucination but you get the right answer?
Tends to latch onto whatever was true at the time of its training run unless you point it to the llm.txt.