Human-in-the-loop: Our core differentiator

Our goal of implementing AI for customer experience is not to entirely replace people. Instead, our AI frees CX professionals the tedious task of resolving simple or repetitive customer support queries and allows them to focus on more complex issues. 

At Crescendo, our team of CX professionals handles escalations when our AI decides that it would be beneficial to include human input. For example, consider a recent conversation our bot had with one of our client’s customers. The customer was having issues using our client’s product and couldn’t figure out how to turn it on. While the bot offered support and expressed its compassion to the customer, it realized that a hand-off to a live agent would be more productive because the customer kept experiencing the same problem and likely required processing of a return. After receiving the bot escalation, one of our agents successfully processed the return and replacement of the product, finishing the conversation with a very strong CSAT score (8/10). 

What would the outcome be if we didn’t use human-in-the-loop? Well, our bot would have tried its best to resolve the issue, and it would likely have gone very far, but the customer would have gotten stuck and had to start-over with a contact center experience.  It would have been significantly more customer effort and resulted in a lower quality customer experience. In other words, escalating to live CX professionals is sometimes necessary for ensuring successful customer outcomes.

While GenAI Assistants form the core of our AI service, we also use AI to review and elevate quality and performance. With our 100% Voice of Customer coverage guarantee, every conversation is scored and each conversation – automated or human – is evaluated by our AI technology to help them ensure a high CX standard.

Every CX professional can see twelve different KPIs that demonstrate how well they handled an individual conversation and how well they’re performing in general. These KPIs are all calculated by our AI using conversation transcripts. Let’s take a look at a snapshot of how four of these KPIs appear on an agent dashboard:

This particular CX Professional has relatively high scores for Sound like an Expert (Exp), Flow and logic of conversation (Flo), Language and tone (Lan), and a lower rating for Net Promoter Score (NPS). Exp measures an agent’s confidence and professionalism, Flo measures the logical flow of the conversation, Lan measures how friendly and helpful they are, and NPS measures how likely the customer is to recommend the service to a friend. Overall, the agent performed pretty well over these 4 categories.

Each individual agent conversation contains these KPIs as well as an explanation for the rank:

While quantitative feedback for improvement is important, our AI also includes personalized, qualitative reviews for an agent’s strengths, weaknesses, and areas for improvement. These recommendations are critical to informing our CX professionals of steps they can take to elevate their performance. 

In conclusion, our team of CX professionals plays an important supporting role to our CX Assistant, delivering an overall outstanding CX service. Our generative AI feedback and KPI scoring lets our professionals know exactly how they performed and what they need to improve to achieve their customer service goals. With human-in-the-loop, we can ensure that our customers’ needs are taken care of with the best CX experience.