Table of Contents
Selecting the right use cases is one of the most important decisions CX leaders face when deploying generative AI agents in customer service. Good use case selection can demonstrate value and ROI, secure internal support, and provide insights that streamline future deployments.
So how do leaders decide where to start? Here’s how they prove value, set priorities, and keep customers at the center in their own words.
Lay the groundwork with the right data
James Dill, Digital & AI Transformation Specialist, Assurant
From AI-enabled to AI agents: A strategic on-ramp for your contact center
I know a lot of companies, ourselves included, tend to look at data as almost a subfactor of performance instead of true performance. There's a lot of bottom dollar type savings, but there are a lot of metrics that inform the rest of the experience that can really help all of those metrics grow as you do. So make sure you have the right data, the right visibility, and the right information. I mean, that's really the groundwork we've been able to look at for any new implementation. Check those boxes, and you're good to go.
With data as the foundation, you can then look at how to prioritize which problems to tackle first.
Filter by containment, complexity, and customer hot-button issues
Director, CX Digital Transformation, Fortune 500 Airline
We had a lot of intents we could look at in terms of data, especially those with lower containment. Then we added the lens of complexity, and finally, what we felt were the most pertinent problem statements from a customer perspective—the hot-button topics. When you create that intersection, the right intents start to fall out. That’s the filtration process we used for the short list you see here.
After narrowing down your focus, the next challenge is how to actually begin deploying AI against those use cases.
Start small and prove it out
Harry Clapham, Director of Operations Strategy and Enablement at Tangerine Bank
Tangerine’s Path to Agentic AI: Building Better, Smarter CX in Banking
It's important when you deploy these technologies, you pick, you know, the right limited use case. Maybe it's an FAQ. Maybe it's a transactional use case. You prove it out, and you expand over time.
What might be smarter is if you can get a goal on the board or get a score on the board, which might be a simpler use case, but you show that it gives the returns that you're saying it will, that everybody's comfortable with the risk. It might be internally facing to begin with, but you're still proving out how generative AI understands human prompts. There's a whole other learning you can obtain that applies to future use cases, but your starting point has been kind of cognizant of the concerns around the table, but you're getting points on the board.
Harry emphasizes here the principle of getting “points on the board.” James reinforces this with specific examples of what that looks like in practice at Assurant.
James Dill, Digital & AI Transformation Specialist, Assurant
From AI-enabled to AI agents: A strategic on-ramp for your contact center
What's the lowest-hanging fruit of queries that customers have for us that can be solved by a generative agent—whether that be a question about their terms and conditions, the status of their claim, or filing a new claim? We think about the low-hanging fruit and the eighty percent of interactions we have that could be addressed via a generative agent. And we've also considered those instances where the generative agent can help get you ninety-nine percent of the way there, but a human still needs to make the final decision.
Alongside data and smart piloting, people play a critical role in identifying the best use cases.
Bring your people with you
Nikki Schmidt, Director of Technology Services, Assurant
From AI-enabled to AI agents: A strategic on-ramp for your contact center
Bring your people with you. Your agents are a wealth of information. Your business partners are a wealth of information. Nobody does not have value in this discussion, and especially those who are on the front lines of your systems, your day-to-day. They're gonna know the best about your data, about your knowledge bases, about what's happening with your consumers.
With the right people and data aligned, the focus turns to what matters most: the customer.
Focus on reducing friction
Steven Canterbury, Director, Customer Success Management, ASAPP
The AI ROI equation: balancing operations, costs, and better CX
The biggest thing I’ve seen from companies doing AI strategically is that they’re focused on reducing customer friction—making it easier for customers to get to a resolution, whether through self-service, faster access to an agent, or shorter queue times. The real goal of delivering AI right now is reducing that friction in the first place.
That focus on friction becomes especially important in moments of disruption, where AI can help manage scale and speed without sacrificing experience.
Turn disruption into faster resolution: an example
Steven Canterbury, Director, Customer Success Management, ASAPP
The AI ROI equation: balancing operations, costs, and better CX
We worked with an airline that delivered an AI solution to identify when customers were impacted by a disruption and let them rebook on their own, with natural interactions. The result: higher containment, faster resolutions, and freed-up agents for complex issues. For disrupted customers, live-agent resolution took about 25 minutes end-to-end. With the generative AI solution, it took just eight.
Why use case selection matters
For CX leaders, selecting use cases isn’t just a starting point. It is an important decision in deploying AI agents. These perspectives show that the best use cases aren’t the flashiest ones, but the ones that reduce friction, prove value, and solve real customer problems. The organizations that succeed are disciplined in how they evaluate, practical in how they deploy, and always grounded in what matters most: improving the customer experience.
.jpg)
