In Part 2, Venice explains how Rockhop redesigned a client’s support process using Copilot Studio and AI to make case handling faster and smarter. The new approach delivers tailored customer responses immediately, assigns cases automatically, and creates knowledge articles on the fly to build a stronger support library. These improvements revealed where the biggest gains—worth $2.5 million—were hiding.
Hey everybody, welcome back to part two of uncovering $4 million of ROI at our enterprise customer. In part one, I shared with you the situation and how we got to the point that we were looking at support's overall process of working cases and what can be improved.
What we proposed for the new support process is that we would use Copilot Studio to be that first receptor for the customer. Making sure that we have a polished tailored response to every single customer and right off the bat giving them the answers that are asked a lot really really quickly. Right? So we are improving the quality of the chat experience right off the bat. From there, we would use the knowledge article base for the co-pilot studio agent so that we can pair the customer with the appropriate knowledge article as it makes sense. Then in the cases that it does need to be escalated because this problem hasn't happened before, that can happen and we would hand it off to one of their support experts.
If a case came in, we would use AI to auto assign the right category. So instead of somebody having to look at the case and say this is dealing with this product, let's put it here, we can use AI to understand the case details and make a first assumption about what category it is. So we're already acting more quickly than we were previously. Then if a knowledge article isn't available, let's use AI to draft that and send that to a human in the loop that can then understand the case, the knowledge article draft, what needs to change, and we can improve it from there and publish. So this is upping the amount of knowledge articles that are created for those cases that do get escalated. So we can start to close the loop of cases that need to be escalated to a human to continually improve our overall process.
There still are situations where a case is created and then it does need to get escalated or it was being worked by somebody and their shift ends and we need to hand it off. The biggest pain point that we saw was that they then need to read the chat history or the email history for this case. This is the place that had the most ROI. In the next video, I'm going to break it down for you of how these changes are resulting in $2.5 million and what we're going to do about it. Hope to see you there.