AI assistant diagnostic:
do you have anything to feed it?
An AI assistant does not fail on the model. It fails because there is no documentation to answer from, because nobody catches the conversation when the assistant does not know, or because nobody looks at what people asked. Six questions to see whether it would work in your case.
The six areas we look at
Volume and repetition
Whether your enquiries resemble each other enough to automate.
Documentation
What material the assistant would answer from, and whether it is current.
Channels
Where people ask you things and whether you know the volume per channel.
Handover to a person
What happens when the assistant does not know the answer.
Languages
How many languages people write in and which you cover worst.
Measurement
Whether you could say, with data, that the assistant achieved something.
Questions about the diagnostic
Can the answer be that I should not do it?
Yes, and that is a useful result. With no repetition and no documentation, an assistant adds a layer to maintain without taking work off anybody.
Do you analyse my site or my documentation?
No. The diagnostic is built from your answers. Reviewing the actual material is something we do on the call, if you want to get there.
Do I have to leave my email?
Not for the score, the breakdown and your main problem. The email is for the full plan.
Will you call me?
Only if you ask.
Shall we look at your real enquiries?
Half an hour to look through a month of your enquiries and say whether an assistant would remove work or just move it elsewhere.
Book 30 minutesTell us your challenge. We'll propose a solution.
No commitment. Within 24 hours, you'll receive a proposal with scope, timeline and budget. No fine print.