By department · Customer service
Repeat questions say something about the product or the process, not about the service desk. By clustering tickets on what is actually being asked, it becomes visible which questions can be solved at the source and which still need a person.
We cluster tickets by content, make sentiment measurable, and decide together which questions lend themselves to an automated answer and which do not.
The work itself is described under generative ai and data analysis.
New Data Student makes visible which questions recur and what they cost: categorise conversations, count them and link them to where they come from. New Data Squad steps in on setting up generative AI on your own knowledge base, including the agreements on what may and may not go in. New Data Search finds the analyst who keeps this up permanently.
Because the cause rarely sits with customer service. A recurring question is almost always a signal from another process: an unclear invoice, a step in the order nobody understands, an email that arrives too late. Making visible which question recurs how often points at that cause.
Part of them, and the part it can handle is usually the dull part. What it cannot do is judge an exception. We set it up so an employee sees the answer and sends it, rather than a bot talking to your customer directly: that saves time without you losing control.
Start with the New Data Scan: one day on site, fixed price, and afterwards you know which questions cost the most and which can be removed at source.
These are indicative figures from earlier work and not a promise. What it delivers for you depends on where you are now.