Why omnichannel is slowing AI adoption in enterprises
You don't have omni-channel. You just have more places to get it wrong.
In this episode of The Experience Edge, Jochem van der Veer opens with a utility company onboarding half a million customers a year. Solid acquisition machine, low churn — and a cost-to-serve problem that was getting out of hand. The root cause wasn't the product or the pricing. It was onboarding. Customers who came in through brokers or over the phone had fragmented profiles, misaligned data, and no continuity across channels. Every support interaction started from scratch. The episode uses that story to make the case that slow AI adoption in enterprise isn't an AI problem. It's an omni-channel problem that nobody fixed before adding the next layer.
Main takeaway: Omni-channel doesn't mean being everywhere. It means customers don't have to repeat themselves. Most enterprises never built that. They added channels. And now AI is inheriting every gap they left behind.
The cost of getting omni-channel wrong
The utility company's data told the story clearly. Customers who onboarded digitally had everything tied to their profile. They could self-serve. Only hard issues reached a human. Cost to serve: low. Customers who came in through brokers or third parties had fragmented profiles, missing data, wrong bills, failed payments. Every support interaction meant rebuilding context from scratch. Cost to serve: much higher.
The difference wasn't the number of channels. It was whether the experience was continuous from day one.
The same gap that made support expensive is the gap AI now has to work with. Drop an agent into that environment and it doesn't solve the fragmentation. It handles conversations confidently and gets them wrong confidently.
What omni-channel actually means
The last decade produced a common misread: omni-channel means being present on every channel. It doesn't. It means a customer can move between channels without losing context. It means they don't have to repeat themselves to your agent, your chatbot, or your AI.
The broken version is easy to recognize:
Customer onboards via broker
Calls support, repeats their full situation
Sends an email, which ignores the call
Visits a store, starts over again
Tries the AI assistant, which has no idea what happened
Multiple touchpoints are not the same as a consistent experience. The channels got added. The design work didn't follow.
What AI inherits
When AI gets dropped into a fragmented environment, it doesn't fix the fragmentation. It automates it. The missing data is still missing. The siloed profiles are still siloed. The KPIs still optimized per channel instead of per journey.
The result: deflection metrics go up while customer satisfaction goes down. AI looks like it's working by one measure and making things worse by the one that actually matters — what does it cost and what does it feel like to be a customer across the full lifecycle?
AI is the most powerful context engine available. But it can only use context that already exists.
Four things that make it work
The utility company is working through four changes to fix the foundation:
Journey-level context. Not who owns the channel, but which teams are part of delivering the end-to-end experience. AI makes this question more urgent because when an agent acts on incomplete context, the consequences move faster than human mistakes.
A shared data layer. Not just integrations, but context that travels with the experience. Journey context — a step-by-step understanding of what customers go through — is what makes AI actually work. Not the model, not the interface. The context it has access to.
Cross-functional KPIs. If each team optimizes its own channel, the end-to-end journey breaks. The number that matters is the full lifecycle. Shared KPIs give every team a reason to contribute to the same outcome, not just protect their own metric.
A common experience language. Teams need a shared way to describe journeys, customer states, and progress. This is also what makes AI trainable. You can't prompt your way to alignment if the underlying understanding of the journey doesn't exist across the organization.
Context has to exist before AI can use it
The utility company didn't need a better call center or smarter AI. They needed a clear picture of where the journey broke — at onboarding — and organizational alignment to fix it.
Once they had a shared KPI like total call minutes per year, the root cause became visible. Unnecessary calls weren't a support problem. They were an onboarding problem. Fix onboarding, teach customers how to self-serve, give them the tools to problem-solve, and the calls stop happening.
Customers don't think in channels. They don't think in AI versus human. They just act according to what they need. If they have to repeat themselves, the foundation isn't there. More channels won't fix it. Better AI won't fix it. Continuity will.
Want to hear the entire conversation? Watch or listen to this episode of The Experience Edge: