Nexus AI Call Coaching

Previously, calls were not routinely transcribed and coaching tended to rely on more generic, occasional training. Important performance information was also held separately in the calling platform. That made it difficult for a busy manager to consistently review conversations, identify recurring issues and tailor support to individual team members.

The opportunity was to make feedback more systematic without requiring someone to listen manually to every call.

I conceived and built the system independently, developing its evaluation approach and dashboards and refining the output through feedback from the Sales Team Manager. The goal was to create a practical management tool, not simply add AI for the sake of it.

Call transcripts are assessed against defined guidelines and evaluation criteria. The system summarises the conversation, identifies relevant outcomes and provides structured scores out of five for the criteria being reviewed.

Individual evaluations feed into aggregated scorecards and dashboards, helping managers see patterns across calls and over time. Calls that may need closer attention can be flagged for review, while suggested coaching actions give managers a starting point for follow-up.

The platform is live and has received a positive response from the team. It has helped make coaching discussions more targeted, and some individual improvements have been observed. I would not present those observations as an independently verified performance uplift or claim that the platform alone caused them.

The useful part of AI is not the ability to produce a score; it is the ability to make a large volume of information easier to act on. The design challenge was turning raw transcripts into feedback that a real manager could use, while keeping the results interpretable and open to review.