Blog & insights from Premium Plus

AI in CX: Why Most Teams Are Still Stuck in Pilot Mode

Written by Agustina Rojo Flaherty | 19 February, 2026

AI is being used almost everywhere.

But most organisations are not seeing full business impact.

According to The State of AI in 2025 (QuantumBlack, AI by McKinsey), 88% of organisations say they use AI in at least one business function. However, most are still in experimentation or pilot phases.

Only 39% report enterprise-level EBIT impact from AI.

At the same time, expectations are rising.

The Zendesk CX Trends 2026 report shows:

  • 83% of consumers believe customer experience should be better than it is today
    85% of CX leaders say customers will leave a brand over unresolved issues, even after a single interaction
  • 74% of consumers expect 24/7 service due to AI

AI is now standard. Customers expect it.

The difference between average and high-performing organisations is not access to AI tools.

It’s their strategy, how they implement and scale them.

Speak with an Expert

 

What the Research Really Shows

 

1. AI Is Broad, But Not Deep

Many teams have:

  • A chatbot
  • Some ticket automation
  • AI assistance for agents

But scaling AI across workflows is much harder.

McKinsey shows that high performers redesign workflows and embed AI into core processes.

This means:

  • Changing how tickets are routed
  • Defining when AI acts vs. when humans validate
  • Aligning AI initiatives with business objectives
  • Having leadership ownership

Most organisations have added AI features.
Fewer have changed how the operation runs.

 

2. Agentic AI Is Growing

62% of organisations are experimenting with AI agents.

Zendesk reports that 87% of CX leaders believe agentic AI can improve interaction quality.

AI agents can:

  • Handle multi-step processes
  • Retrieve knowledge
  • Make decisions like refunds or returns
  • Work across channels

But they depend on:

  • Clean data
  • Structured knowledge
  • Clear governance
  • Strong integrations

Without that, AI increases complexity instead of reducing it.

 

3. Context and Memory Now Matter

Zendesk highlights “memory-rich AI” and contextual intelligence.

Customers do not want to repeat themselves…

  • 74% find repetition frustrating
  • 85% of CX leaders say memory-rich AI is key to personalised journeys

For decision-makers, this translates to practical questions:

  • Is your knowledge centralised?
  • Do agents see full customer history in one place?
  • Does AI retrieve accurate information?
  • Are systems properly integrated?

If the answer is no, AI personalisation will not work as expected.

 

“Turning on AI agents in Zendesk is easy. Making them reliable, accurate and aligned with your processes is the real work. That’s where redesign and governance become critical.”

Benoit Smagghe, Solutions Consultant Premium Plus

 

4. Speed Is Expected. Resolution Is What Counts.

AI has raised the bar.

Customers now expect:

  • 24/7 availability
  • Fast responses
  • Accurate first-contact resolution

High-performing organisations do not focus only on efficiency. They set goals around growth and innovation as well

For CX and Operations leaders, this means:

Move beyond:

  • First reply time
  • Ticket volume reduction

Focus on:

  • First contact resolution
  • Containment quality
  • CSAT
  • Cost per resolution
  • Value-based: Quality of reply

AI success is not about automation volume.
It is about business results.

 

Premium Plus Perspective

We see the same pattern repeatedly.

Most organisations do not struggle because they lack AI.

They struggle because their operational foundations are not ready.

Common issues we see:

  • Over-customised Zendesk instances
  • Old workflows layered over new automation
  • Knowledge bases not structured for AI retrieval
  • Reporting that does not reflect real performance
  • AI enabled without governance

Turning AI features on does not equal transformation.

If workflows, data and knowledge are not structured correctly, AI exposes weaknesses instead of solving them.

Our view is simple:

Before scaling AI, make sure your Zendesk setup, workflows and knowledge architecture are solid.

Otherwise, you risk adding cost and complexity.

 

How Premium Plus Helps

As a Zendesk Premier Partner, we work with CX, Support, IT and Operations leaders across EMEA to make AI practical and measurable.

 

1. Zendesk Health Check

We review:

  • Workflow design
  • Automation logic
  • Knowledge structure
  • AI configuration
  • Reporting setup

We identify gaps and redesign for scalability.

 

2. Workflow Redesign for AI

Based on what high-performing organisations do, we help you:

  • Embed AI into real operational processes
  • Define validation and governance
  • Align AI initiatives with business outcomes

 

3. Knowledge & Help Centre Optimisation

Contextual intelligence depends on strong knowledge foundations

We:

  • Structure content for AI retrieval
  • Reduce duplication
  • Improve self-service accuracy

 

4. Integrations & Data Flow

AI cannot work in silos.

We connect Zendesk with your CRM, ERP and internal systems to ensure:

  • Context flows correctly
  • Agents have full visibility
  • AI decisions are informed

 

5. Training & Adoption

AI adoption requires people alignment.

We support:

  • Agent training
  • Admin enablement
  • Leadership alignment
  • Governance frameworks

 

Download the CX Trends 2026 Guide

To explore the five trends shaping customer experience, download the CX Trends 2026 guide here:

 

 
 

It gives practical insight into where customer expectations are heading.

 

 

If You’re Reviewing Your AI Strategy

Ask yourself:

  • Are we still in pilot mode?
  • Have we redesigned workflows, or just added tools?
  • Is our knowledge AI-ready?
  • Do we measure resolution properly?
  • Do we have clear ownership?

If the answers are unclear, it may be time to review your setup.

Contact Premium Plus to learn how we can help you scale Zendesk and AI in a structured, measurable way.