One of the changes I'm most proud of this past year was leading the implementation and evolution of Agent Copilot combined with AI Scoring (Virtual Supervisor) in high-volume operations.
At the beginning, the challenge wasn't just deploying the technology, it was driving real adoption and making it part of the agents' daily workflow.
So I shifted the focus from just technical delivery to the actual experience. I stayed very close to the agents, collecting feedback constantly and using that to refine Copilot prompts like summaries, resolutions and next steps based on real conversations. At the same time, I built custom reports to track usage and behavior, and introduced some gamification strategies to increase engagement. We also aligned AI Scoring with operational goals to reduce the manual effort in quality.
What makes this stand out to me is that it wasn't just about implementing AI, it was about making it truly useful and trusted by the people using it every day.
As a result, we saw a strong reduction in AHT, with chat operations going from around 15 minutes to about 9. We also reduced QA time from around 18 minutes to about 4.5 minutes per evaluation, while increasing consistency and scalability. On top of that, agents started to rely more on Copilot and feel more confident during interactions.
In the end, we improved both sides of the experience. Customers benefited from faster and more consistent interactions, and agents had less cognitive load and clearer guidance during their work.
That's what makes this one special for me. It wasn't just a tech project, it was a real transformation in how people work and deliver experience.
------------------------------
Mateus Nunes
Tech Leader Of CX at Solve4ME
Brazil
------------------------------