The interview
Before Elmtrail, how did you feel about bringing AI into your team’s work?
Working in support operations, I see every day how much of the team’s energy goes into work that repeats itself, and how little is left for the cases that really need a person. I’ve always been excited about what AI could do here, and just as curious about how to apply it to real customers without losing our quality.
What was the moment you realised it was actually working?
After a couple of weeks spent training Elmtrail on our knowledge base, more than half of our customers’ questions were being answered by the AI on its own, quoting our policies and FAQs correctly and cleanly. And the customers kept writing back to Mario, our agent, as if he were a real colleague. That was the moment it stopped being an experiment and became part of the team.
How does a typical day look different now?
Far more organised. Instead of chasing the queue, we look at the whole picture: where the requests come from and what keeps coming back. Our attention goes to the conversations where experience really matters, and my own day has shifted from checking tickets to improving how we work: what to add to the knowledge base, which flow to build next.
What has Elmtrail made possible that wasn’t possible before?
Freeing the team from the repetitive part of the job while keeping full control: every answer is built on our own knowledge, every step is visible, and we step in wherever a decision needs a person. And improving fast: when a rule changes, we update our knowledge base once, test it on real cases, and from that moment on everyone answers correctly. For a support team, that is a dream come true.
What would you say to a team on the fence?
Go for it, and start with the questions you answer most. Look at what it prepares, approve, correct where needed. The first time your team recognises its own quality in the answers, and realises how much time it just got back, the decision makes itself.
