We lead our teams and serve our clients with our core values of: Make an Impact, 100% Responsibility 0% Excuses, and Grow or Die. When we began exploring how AI could enhance customer service for our team and our clients, it was essential that we stay focused on our core values.
Leading with Core Values in AI For Customer Service
AI can create massive efficiencies. It can also create distance. We made a deliberate choice to avoid creating that distance. There are AI tools that could distance our engineers and force our clients to text and talk to AI before an engineer. But that would feel more like customer avoidance than customer service. It doesn鈥檛 reflect our values, and more importantly, it doesn鈥檛 provide our clients the experience they deserve.
Building Magic Agent: Testing, Transparency, and Trust
When we rolled out our new chat tool, Thread, we spent six months in testing because there鈥檚 a significant amount of intelligence required behind the scenes. Building Magic Agent, our AI support assistant within Thread, took ten months of testing, rebuilding, piloting, monitoring, and tweaking. Those tweaks are still happening today, and all are being led by humans focused on customer service first.
Before we drafted a single prompt, we asked one question: What is the impact on our people and our clients?
We debated what 鈥�100% Responsibility鈥� means when an AI receives negative feedback. Ultimately, the responsibility falls on us. Every piece of feedback is received by our continuous improvement team, and we act on it.
We examined what it means for AI to 鈥淢ake an Impact.鈥� Within minutes of launching our chat tool, our response-time metric dropped to eight seconds. Impressive, but not an honest reflection of how our clients measure our service. AI forced us to rethink what truly matters, so we shifted our focus to other, more meaningful metrics.
We made design decisions through the lens of our values. Your ticket enters our engineering queue the moment it鈥檚 created, even if you鈥檙e still chatting with the bot. An engineer can jump in mid-conversation, and that鈥檚 by design. If you鈥檇 rather skip the bot entirely, you can ignore it or say 鈥渆nd,鈥� 鈥渆ngineer,鈥� or 鈥減erson,鈥� and it steps aside.
We are transparent by choice. In your chat window, it says 鈥溩畲笞试床杉� Thread (bot),鈥� because we want you to know exactly how you are being helped.
Measuring Impact: Beyond Speed to Meaningful Metrics
We encourage you to use the bot, because it can save you time, not so we can hide behind it. Last week, 8.5 percent of users who engaged with Magic Agent resolved their issue without needing an engineer’s assistance.
Improved service doesn鈥檛 happen accidentally. We review internal feedback, user sentiment, CSAT, and your direct feedback. All the feedback comes to my team and serves as the catalyst for further improvement.
Our goal is simple: deliver the best service possible. That was true before AI, and it鈥檚 even more important now.
If we鈥檝e missed the mark, I want to hear from you. If you loved your experience with our chatbot or our engineers, I want to hear that too.
And if you鈥檙e ready for an exceptional IT team with AI-enhanced service, let鈥檚 talk.
