The quarter just ended. Three numbers worth looking at
Not a strategy review. Three figures you can pull in twenty minutes that tell you whether your AI spending is doing anything.
The third quarter closed on Tuesday. Before the figures get folded into a year-end report nobody reads closely, three numbers are worth pulling out and looking at on their own.
None of them require a project. All three come from invoices and a short conversation.
Number one: what you actually spent
Not the subscription you remember. The total, including AI features bundled into software you were already paying for, which the 15 September post noted is usually the larger figure and almost never counted.
Most companies doing this for the first time find the number is either much smaller than they feared or spread across four invoices they had never added together. Either result is useful.
Number two: how many people used it
Most tools report active users to an administrator. Look at the count, then compare it with how many licences you are paying for.
If you are paying for twenty and six people are active, you have the problem from 14 August, and the fix is not more training — it is asking five of the fourteen what happened.
Number three: one process that changed
This one is a question rather than a figure. Name one thing that is genuinely done differently now than in June.
If you can name one, the spending is working and the number in column one is irrelevant. If you cannot name any, then whatever you spent bought optionality rather than results, and that is worth knowing before you renew.
What to do with the three
Write them in one place with today’s date, and do it again on 2 January. The comparison between two quarters is worth more than either quarter alone, and almost nobody keeps the first measurement.
Twenty minutes, once a quarter. It is the cheapest management instrument available in this area and it replaces a great deal of discussion.