15/10/2026
A founder told us last month that his new AI chatbot was ""broken."" π€
It was quoting customers the wrong prices. Confidently. Every single time.
We looked under the hood, and the AI was working perfectly. The problem was what it was reading. Product prices lived in three different spreadsheets, and two of them had not been updated since last year.
The chatbot was not broken. It was simply repeating the mess it had been given, at speed. π
This happens far more often than people admit. Businesses rush to add AI, then blame the tool when results go sideways. The real issue usually sits one layer below.
So before you spend a dollar on AI, ask your team:
ποΈ Where does our most important data actually live?
π When was it last checked for errors and duplicates?
π€ Do all our systems agree on the same numbers?
If those questions make the room go quiet, you have found your starting point. Fix the data first, and AI becomes a genuine advantage instead of an expensive amplifier of old mistakes.
Smart AI starts with honest data. π§±
Want the questions we use to assess AI readiness? Comment ""READY"" and we will send them over. π