A recent AI experiment involved inputting handwritten puppy litter box logs into ChatGPT over several weeks. My wife and I tracked every pee, poop, walk, and accident of our puppy, Oliver, and created a new kind of KPI: a key potty metric.
ChatGPT extracted the data from your scribbled notes and converted them into fully formatted CSV files. Next, we went to town to actually graph these KPIs.
Many fancy metrics have been devised, including ARR (accident reduction rate), LTV (longest toilet time), PTP (poo to pee) ratio, WAF (weekly accident free) rate, and DPV (daily toilet volume).
This graph was helpful because it showed how accidents correlated with the frequency of Oliver’s walks.
Graph showing trends in puppy potty training data Alistair Barr/ChatGPT
As a result of AI analysis, it was determined that the accident was not a coincidence. They are concentrated during two predictable times: midday (approximately noon to 3 p.m.) and evening (8 p.m. to 10 p.m.), often after a meal, nap, or play session.
Biggest takeaway: Success comes from reducing the gap between “must go” and “opportunity to go.” According to Oliver’s KPI dashboard, consistency trumps discipline, every accident is a training opportunity, and our pup is becoming increasingly predictable.
Sign up for BI’s Tech Memo newsletter here. Please contact us by email. abarr@businessinsider.com.