The Experience Edge: Why AI needs journey context to actually work with the Insitute for Journey Management

Mark Smith and Raymond Gerber have spent a combined six decades building competing visions for the future of journey orchestration: Mark at Kitewheel, Ray at Thunderhead. They weren't just rivals in the same market; they were making fundamentally different bets on how enterprise CX would evolve. Now, as co-founders of the Institute for Journey Management, they're on the same side of a more urgent question: as AI floods into the enterprise, what actually makes it work?

In this episode, Jochem van der Veer sits down with both of them to find out.

Here's why you don't want to miss this episode

Journey context isn't a feature. It's the foundation

Before the conversation goes anywhere, Ray draws a line in the sand: journey context needs a proper definition, because the industry has been fuzzy about it for too long. He breaks it into four dimensions: temporal (context changes over time), situational (the same customer needs a different response at onboarding versus renewal), directional (context should point toward an objective), and constraint-based (it must reflect what the business can and cannot actually do). That last dimension is the one most AI implementations get wrong.

Mark builds on this from a data science angle: algorithms will learn whatever is in your data. If your data is full of mediocre interactions, your AI learns mediocrity and then delivers it at scale. Journey context is the mechanism for teaching AI what "good" looks like, grounding optimization in business reality rather than statistical convenience.

The real AI breakthrough isn't prediction. It's prescription.

Predictive models have existed for decades. Mark's been building them for 35 years. They work reasonably well, around 70 to 80 percent of the time, but customers are unpredictable and the world changes. The genuinely new development, Ray argues, isn't better prediction. It's the collapse of the gap between discovery and execution.

Generative AI, combined with graph-based retrieval and the fusion of structured and unstructured data, creates something that didn't exist before: a closed feedback loop that continuously reassesses customer intent and updates the recommended next conversation in near real-time. You're no longer running a prescriptive model once a month. The system is learning and adapting continuously. For organizations that have been stuck in batch-mode CX, this is a fundamental shift, not just in capability, but in how quickly you can respond to customers whose behavior is changing faster than your models ever could.

Hyper-personalization finally means something

Ray makes a distinction that's easy to miss but impossible to unsee once you catch it. The old version of "hyper-personalization" was really just sophisticated segmentation: 500 next-best-action templates fired at 12 million customers. That's not one-to-one. It's one-to-cluster.

What generative AI unlocks is something closer to the genuine article: taking a next-best action recommendation and generating a truly individualized communication around it, shaped by the specific customer's language, history, location, and context. The number of possible outputs becomes effectively infinite. Ray draws a direct line from this to changing consumer expectations. Customers who interact with Claude or ChatGPT are developing a new baseline for what "helpful" feels like, and they're going to expect the same from their bank, their insurer, their utility provider. Organizations that are still firing template emails are going to feel that gap very soon.

Journey context matters more inside the business than outside it

This is the most counterintuitive claim in the episode, and Mark makes it without flinching: the customer just lives their journey. They don't need to understand it. But inside most organizations, journey thinking is deeply unnatural. Teams are organized by function, rewarded on departmental metrics, and have no structural reason to care about what happens two steps downstream from their touchpoint.

Jochem illustrates this with a utilities company that had significant call center costs driven by billing friction. The root cause wasn't billing. It was onboarding: specifically, that customers who weren't set up on digital channels at the start ended up generating a disproportionate volume of support contacts throughout their entire lifecycle. The billing team couldn't see that. The onboarding team didn't know they were causing it. Journey context is what makes that connection visible and actionable for the people who have the power to change it.

From journey mapping to journey-led operating model

Both Mark and Ray are direct about one of the industry's persistent failures: journey initiatives have been positioned as projects rather than operating models. You run a discovery sprint, you build a beautiful map, you present it to leadership, and then it sits in a folder somewhere while teams continue optimizing their individual KPIs.

Ray's example of a German utilities company with 20 million customers illustrates what success actually looks like. The difference wasn't technology or methodology. It was that a senior director with both accountability and authority ran every engagement around the end-to-end journey (not isolated touchpoints) and moved the organization away from departmental metrics toward journey-level outcomes. Everyone was rewarded based on what the journey produced, not what their team produced. That's an operating model change, not a CX project.

Mark's example of Synchrony Financial points in a similar direction. With 90 million customers across branded card products for Amazon, Best Buy, and Walmart, Synchrony has almost no direct brand relationship with the people it serves. The experience is the product. What made their journey practice work was a centralized analytics leader who sat above the divisional structure and could connect the dots that individual product teams couldn't see.

The CX function's identity crisis—and its opportunity

Neither guest is sentimental about traditional CX teams. Mark's assessment is pointed: too many CX groups have forgotten they're supposed to care about customers. They've become survey machinery: measuring NPS, managing VOC programs, reporting satisfaction rates, without anyone asking why those numbers are what they are or who has the authority to change the underlying causes.

The opportunity, Mark argues, is for CX leaders to step into the connective tissue role—the function that gets above the silos, understands how decisions in one part of the business land on customers in another, and orchestrates action across teams that have no other reason to collaborate. Ray adds that this only works if CX stops being an isolated function and becomes embedded in how the organization operates, with the CTO and CIO taking real ownership of the enabling infrastructure.

The question is whether CX leaders will grab that mission before someone else does.

The conversation that CX leaders need to have right now

What Mark and Ray bring to this episode isn't a technology pitch. It's something harder to come by: pattern recognition from two people who have watched the same cycle repeat across three decades of enterprise CX evolution—the promise of scale, the failure of implementation, and the gradual realization that the bottleneck was never the algorithm.

The bottleneck is always organizational. AI doesn't change that. It makes it more visible, more urgent, and for organizations willing to do the work, more tractable.

If you're building an AI-enabled CX practice and wondering why it's not delivering what it promised, this is the episode that will help you understand why.

Want to hear the entire conversation? Watch or listen to this episode of The Experience Edge:

Spotify | YouTube | Apple Podcasts