In my first testing of OpenAI's ChatGPT3 product, I threw it a softball question and was completely surprised by its replies that it made me doubt anything else it might have to say.
My first question was simple enough: Who is Michael McCafferty?
Of course there are many people with my name so I wasn't so surprised when ChatGPT came up with a different one for its first answer... check it out:
So I took a different subject, one which I am personally familiar with, the subject of TeleMagic software:
It knows what TeleMagic is and it knows that it is a CRM product, but it thinks TeleMagic was created in 1983, four years before Act! which it thinks is the first CRM (1987... which is what it says on the Act! page in Wikipedia). There's some very basic logic that is missing here. ChatGPT should know better.
Notice that NexCustomer is not on the list! Yikes. For answers I turn to a real human, Rich Bohn, who is the "oldest living independent CRM analyst". Rich has never heard of either Doni Davenport or NexCustomer. Showing no fear, I press on:
Needless to say, the link for Donny D. is broken, no info there. You might think that ChatGPT would check before just running off at the mouth like that, but noooo.
For the final insult, ChatGPT gives 3 different answers to the same question. This is where I stopped. My brain was fried.
For reference, here is the list of the earliest known CRM products and links to sources: