There is no shortage of conversation about AI right now. What interests us at Referential is what happens after the conversation. What does AI actually look like inside a customer advocacy program? Where does it genuinely make the work better? And, perhaps most importantly for a discipline built around relationships, what should we be asking AI to do, and where does human expertise become even more valuable?
We have had the opportunity to work alongside several of our clients as they begin answering those questions in very practical and very different ways. A large enterprise technology company with a sophisticated customer advocacy program is using AI to bring more structure and repeatability to advocacy campaign design. Planview is using it to reduce time spent searching, organizing and producing so people can spend more time thinking strategically and building customer relationships. NiCE is creating a forum where individuals putting AI to work can talk candidly about what they are learning along the way.
Taken together, their approaches suggest something important. AI is certainly changing how advocacy work gets done, but it is also creating new opportunities for advocacy teams to do what they have always done well: connect expertise, experience and people.
Turning advocacy ideas into repeatable workflows with AI
One of our clients, the large enterprise technology company that we referenced above, started with a deceptively simple idea: when someone finishes an expert certification, invite them to leave a review.
Straightforward enough, until you start building it. Anyone who has designed an advocacy campaign knows how quickly a simple idea becomes more complicated. Who qualifies? What if the customer has already engaged in another campaign? How do you avoid contacting advocates too frequently? What happens after the review request? How is external confirmation captured? How do you log participation? And how do you document all of this clearly enough that a technical team can actually build it?
The team used Claude as a collaborative design partner to work through those questions. They mapped the logic conversationally first, moving from the initial trigger through qualifying checks, campaign enrollment, external confirmation and reward logging. This gave them an opportunity to think through different pathways, define what counted as previous engagement and consider how to avoid contacting the same customers too frequently before turning the concept into a formal specification.
They then used AI to help structure the output for two different audiences. A campaign overview explained what the program was intended to accomplish, while a technical appendix covered webhook triggers, automation steps and required data fields. Through additional refinement, the team standardized terminology and created a visual representation of the workflow.
What is particularly interesting about this example is that the value extends beyond the individual campaign. The process also produced a reusable template for defining future campaigns. Instead of starting with a blank page every time someone has a new advocacy idea, the team now has a structured way to work through the questions required to turn that idea into something an implementation team can build.
AI did not come up with the advocacy strategy. The team brought the expertise and context about their customers, the experience they wanted to create and the business rules that needed to guide the campaign. AI helped them explore that thinking, organize it and translate it into something executable. In that sense, this client is using AI for more than efficiency. They are using it to make good advocacy thinking more repeatable, which becomes increasingly valuable as a program and its number of campaigns grow.
At Planview, AI is making more room for the human work
Planview offers another compelling example because it challenges the idea that efficiency and relationships somehow sit on opposite sides of the AI conversation.
Planview has made AI adoption a strategic initiative, investing not only in Glean but also in training employees to use it effectively. Teams are encouraged to experiment, document what they learn and share their ideas with one another. James Lustenader, who leads customer advocacy work at Planview, describes himself as someone who began as an AI skeptic. His perspective has changed through experience, as he has seen what the technology can do and how quickly it can help him accomplish work that previously consumed a significant amount of his time.
One of the first places he saw an opportunity was also one of the least glamorous parts of advocacy: answering a steady stream of questions from across the business. Those questions might involve finding a customer for a particular need, locating a relevant case study or identifying someone with experience using a specific product. The information existed, but finding it repeatedly for people across the organization consumed a significant amount of time.
Every Planviewer has access to Glean, and so the Advocacy team has built an agent that employees could use to query two distinct sources; existing public case studies and Planview’s SlapFive database. Sales could find more of what they needed themselves, and the volume of questions coming directly to James dropped dramatically.
This is important because answering internal requests may be necessary work, but it is not necessarily the highest value use of an advocacy professional’s time. When AI helps surface information that already exists, James can spend more time on customer recruitment calls and more strategic work.
We see the same benefit firsthand. Referential supports sales reference requests within the Planview program. AI can help our consultant search for potential customer matches more efficiently, which gives him more time to evaluate whether those matches are actually right and to focus on the outreach itself. The value is not simply that the search happens faster. It is that more of his time can be spent applying judgments, making effective matches, and engaging with customers.
James shared another example that brings this idea to life. While preparing a customer executive for a conference session, he used AI to support the production work surrounding their conversations. Instead of dividing his attention between listening, taking notes and thinking about how the discussion would translate into slides, he could focus on being present with the customer.
Afterward, he used AI to help turn the recorded conversation into text, presentation content and visuals. During the process, something unexpected happened. James and the customer had previously connected over books, and the customer had mentioned several titles he recommended to peer leaders. AI expanded that initial list with additional relevant suggestions, including a book written by a former colleague of the customer. That AI generated suggestion sparked another genuine point of connection, and the story ultimately became part of the customer’s talk track.
AI did not create that relationship, and it did not understand why that particular connection mattered in the way James did. It created more space and possibility, while James recognized what was meaningful and knew what to do with it.
For advocacy leaders, that is an important distinction. The opportunity is not simply to divide tasks between AI and people. It is to understand where AI can reduce the work surrounding a relationship so people have more capacity to apply the judgment, curiosity and empathy that make the relationship valuable in the first place.
At NiCE, advocacy is creating a space for the AI conversation itself
Like Qualtrics and Planview, NiCE is actively using AI within its own organization. There are plenty of internal applications we could explore, but for this conversation we wanted to focus on something different: how NiCE is using advocacy and content to bring the AI conversation to its customers and the broader CX community.
Its AI focused podcast, Speak to an Agent, brings together the innovators, thought leaders and decision makers shaping the future of customer experience AI. Importantly, the premise is not simply to celebrate what the technology can do. The conversations are designed to get candid about what is working, what is not and what organizations are learning as AI becomes capable of reasoning, planning and acting in increasingly sophisticated ways.
The questions NiCE is exploring move beyond the technology itself. What does it really take to bring AI into a large organization? Why can AI not fix a problem the business has not already solved? Why does real understanding matter more than a polished voice? And what happens to the value of domain expertise when AI becomes widely available?
That last question stands out to us. One of the ideas at the heart of Speak to an Agent is that the domain expert who knows how to use AI is becoming increasingly valuable. That is an important counterpoint to some of the anxiety surrounding AI adoption. Technology may be able to do more of the work, but knowing the business, understanding the customer, asking the right questions and recognizing whether an answer actually makes sense still matter enormously.
There is also an advocacy lesson here. Some of the most valuable thought leadership does not come from a brand explaining what it knows. It comes from creating the conditions for knowledgeable people to share what they know.
By giving practitioners, innovators and decision makers a place to talk openly about their experiences with AI, NiCE is doing more than publishing content. It is convening a conversation around a topic its customers and the broader CX community are actively trying to navigate. From our perspective at Referential, that is an important part of advocacy too. It means bringing credible voices together, giving their experiences a platform and helping an audience learn from the people who are actually doing the work.
Three approaches, with human expertise at the center
These three examples illustrate different approaches to AI, but there is a connection between them. Each puts human expertise at the center.
In the first example, advocacy practitioners bring the strategy and use AI to help turn their thinking into structured, repeatable workflows. At Planview, AI helps remove time consuming work around information retrieval and content production, creating more capacity for strategy, judgment and customer relationships. At NiCE, the focus shifts to the expertise itself by creating a platform where people can share what they are learning about putting AI to work in the real world.
For advocacy leaders, that feels like a more useful starting point than asking how much of our work AI can automate. Instead, we can ask where AI can remove friction, where it can help us apply our expertise more consistently, and where it creates an opportunity to elevate the expertise of our customers and communities.
There is a leadership component to this as well. Planview’s experience demonstrates the importance of giving employees tools and training while also giving practitioners room to determine how AI can best support their individual roles. Teams are encouraged to experiment and share what they learn, creating an environment where one person’s use case can spark another person’s idea.
That mindset matters in customer advocacy because our work has always depended on judgment. We need to know which customer is right for which opportunity, understand when to ask and when not to, recognize the difference between a good story and a meaningful one, and build trust across Sales, Marketing, Customer Success and, most importantly, with the customers themselves.
What we are seeing across these client examples is not a story about AI making that expertise less important. If anything, these examples suggest that expertise becomes more important as the technology becomes more capable. The people who understand their customers, their business and their discipline are the ones who can determine where AI belongs, where it does not and how to turn its capabilities into something genuinely useful.
For advocacy teams, perhaps that is the more interesting AI conversation to be having. Instead of asking what AI can do instead of us, we should be asking what we can do better because we have it.

