New industry perspective highlights the hidden risk undermining AI adoption in CX teams
Many CX leaders are discovering an uncomfortable truth: the biggest barrier to AI success isn’t the model, it’s the foundation beneath it.
In an AI-first CX environment, teams need a responsible way to replatform. That’s why Kustomer built the AI Setup Assistant – designed to help teams configure AI without guessing or relying on brittle, manual setup. AI Setup Assistant connects directly to Zendesk, analyzes an organization’s existing configuration, and intelligently translates it into a new Kustomer environment — surfacing risks, preserving critical logic, and guiding teams through migration with confidence.
“The industry has been asking whether AI is ready for CX,” said Brad Birnbaum, CEO of Kustomer. “What we’re seeing is that AI is ready but many systems aren’t. AI Setup Assistant is about helping teams build the foundation AI needs to perform reliably. Because when AI fails, it’s rarely the model. It’s the configuration underneath it. We believe configuration intelligence will become a core requirement for every AI-first CX organization.”
Across high-growth companies and enterprise CX organizations alike, AI initiatives are stalling not because the technology is immature, but because underlying system configuration quietly breaks at scale. Routing, permissions, automation logic, and data models, long treated as backend “admin work”, are now determining whether AI builds trust or destroys it.
This pattern is showing up consistently where experienced operators are sharing hard-won lessons from AI rollouts that didn’t go as planned.
The AI Myth Holding Teams Back
For years, AI challenges in CX have been framed as model problems: hallucinations, poor accuracy, or lack of sophistication. But in practice, many failures stem from something more mundane, and more dangerous.
AI systems amplify whatever foundation they’re given. When routing is misaligned, permissions are incomplete, or automation fires out of order, AI doesn’t quietly fail. It produces decisive, inconsistent outcomes that erode trust with agents, customers, and leadership.
As one CX leader put it: “AI didn’t fail — our system did.”
Marketing Technology News: MarTech Interview with Stephen Howard-Sarin, MD of Retail Media, Americas @ Criteo
The Rise of Invisible Setup Debt
Modern CX platforms are no longer simple ticketing tools. They are deeply interconnected systems where one configuration decision can affect multiple downstream outcomes.
The problem is that this complexity compounds invisibly. Setup debt builds quietly over time, often introduced by well-intentioned changes made under pressure, a new queue, a new automation, a quick permissions tweak.
Teams don’t feel the impact immediately. They feel it weeks or months later, during volume spikes, migrations, or AI launches, when diagnosing root cause is hardest and stakes are highest.
Why AI Raises the Cost of Getting This Wrong
Before AI, configuration issues were painful but survivable. Teams could patch problems manually, explain away inconsistencies, or absorb inefficiencies with headcount.
AI changes that equation.
When AI underperforms, leadership rarely attributes the failure to setup. The conclusion is simpler, and more damaging: “AI doesn’t work for us.” Momentum stalls, confidence drops, and future investment becomes harder to justify.
In an AI-first world, a clean setup is no longer operational hygiene. It’s a prerequisite for credibility.
What’s often missing from this conversation is that CX setup now serves two workforces, not one.
Historically, configuration decisions were made to support human agents, ensuring the right tickets landed in the right queues, with the right tools and permissions. But as AI agents become part of the CX stack, those same setup decisions – especially the data model choices to ensure well-structured, well-described data in your system – increasingly determine whether AI can act responsibly, accurately, and consistently.
A well-configured CX platform doesn’t just help human reps do their jobs better; it provides the structure, data and context AI agents rely on to reason correctly. Even for teams not deploying AI agents today, the foundation they build now will directly shape what AI can, and can’t, do when they’re ready.
This shift is exactly why the training, tools and processes teams have historically relied on to manage setup risk are no longer sufficient.
Marketing Technology News: From MarTech Stack to MarTech Fabric: Weaving Brand, Content, and Conversion Into One Thread
Why Traditional Fixes No Longer Scale
Many teams attempt to manage setup risk with tools and processes designed for a simpler era:
- Static checklists that age instantly
- Expensive audits or professional services reviews
- Automation tools that auto-fix without explaining impact
These approaches fail for the same reason: they don’t keep pace with systems that are constantly changing.
Worse, black-box automation introduces a new problem: fear. Admins remain accountable for outcomes, but lose confidence when they don’t fully understand what’s happening under the hood.
What High-Confidence Teams Do Differently
The CX teams navigating AI successfully aren’t moving slower. They’re moving with fewer surprises.
- They validate before they automate.
- They look at systems, not isolated settings.
- They keep humans accountable, but supported.
Instead of reacting to failures after launch, they invest in understanding configuration risk before it becomes customer-facing.
The New Requirement: Configuration Intelligence
What’s emerging is a new category of capability: configuration intelligence.
Not automation that takes control away from teams, but setup intelligence that:
- Understands how settings interact
- Surfaces second- and third-order risk
- Explains impact in plain language
- Keeps humans firmly in control
In an AI-first CX environment, teams need a responsible way to replatform. That’s why we built the AI setup assistant at Kustomer. Designed to help teams configure AI without guessing or relying on brittle, manual setup.
From Setup as Admin Work to Setup as Strategy
The shift is already underway. Setup is no longer a backend concern delegated to a few experts. It’s a strategic capability that determines whether AI delivers on its promise or becomes another stalled initiative.
This is the exact problem Kustomer built AI Setup Assistant to address: helping CX teams surface hidden risk, understand system-level impact, and build confidence in their foundation before AI and scale expose the cracks.
Not by replacing human judgment, but by supporting it.
