Does your company know what employees put into ai tools?
Login counts measure adoption, but few companies can say what ‘approved’ AI use actually means
You've probably sat in a version of this meeting. A CEO opens by celebrating adoption numbers, 90% of employees logged into a new generative AI tool within a month.
Twenty minutes later, someone from legal asks almost casually whether anyone knows what employees are actually typing into it. Silence follows. Writing in Forbes, Teramind CRO Ori said he's now watched this exact scene play out at a dozen companies.
Leadership teams have gotten efficient at deploying AI tools quickly, but that speed rarely comes paired with a clear answer about what those tools should and shouldn't be used for. According to the Forbes piece, that gap not the technology itself is where risk quietly accumulates.
Ask five people on the same executive team what "approved" usage actually means, and expect five different answers.
Two years back, internal AI was not something that was even discussed in a sales discussion. Today, it is the subject of almost every discussion and sometimes even brought up by procurement and legal departments because of some probing questions by their customers.
The dashboard on AI usage is very neat to show in front of a board and thus has become a de facto measure of the state of things. However, the number of logins is no measure of whether there is any proper analysis of whether using a chatbot to summarise a client agreement is good or uploading a list of employee salaries is bad.
"Usage is up" gets applause; "we haven't decided our policy yet" gets an awkward silence, which is exactly why the harder conversation keeps getting postponed.
Ori’s recommendation is carefully low-key: one person gets to make the decisions regarding the deployment of these AI tools, not in name but because he or she is the person who is going to be consulted. The question “What can we load into this tool?” has a succinct, clearly worded response available somewhere.
Ori is not advocating that prudence should trump speed. It is clear that companies attempting to escape from AI altogether will have an even more difficult conversation with their board of directors within a year.
But speed without accountability is just momentum, and momentum must hit something sooner or later, writes Ori.
