Yes, we’re getting ahead of ourselves: AI in customer engagement channels

Yes, we’re getting ahead of ourselves: AI in customer engagement channels.

---

Yes, we’re getting ahead of ourselves. The industry is running faster than its own foundations can hold. Everyone is talking about AI – generative models, conversational bots, predictive routing, virtual assistants – and it feels like if you’re not “doing AI”, you’re already behind. But let’s pause for a moment and ask a harder question: are we actually ready for it?

Because if the statistics are anything to go by, the answer is no.

A recent survey across global contact centres showed that only 11% of organisations are not considering AI in the short to near term. Not surprising, given the relentless rhetoric. What is surprising – and frankly, concerning – is that only 48% have even implemented process automation in their contact centre operations, and just 42.9% have achieved workstation consolidation. These are basic structural enablers. Without them, AI doesn’t have a stable foundation to work from.

So yes, I’ll say it plainly: the AI tail is wagging the contact centre dog.

Since we are being honest with each other; most AI roll-outs in customer engagement are still experiments (even though we would like to think of them as strategy).

We’ve all seen the numbers: depending on whose research you read, between 73% and 90% of AI implementations either fail outright or don’t deliver on their promises. In many cases, it’s not because the technology is flawed but rather that the organisation isn’t ready for it. The data is fragmented, the processes inconsistent and the customer journey undocumented. Yet the marketing decks are full of “AI-first” promises that can’t survive contact with operational reality. 

When you put advanced technology on top of weak process design, you’ll get expensive chaos.

Here’s a simple rule of thumb: if your agents are toggling between six screens to help a customer, you are not ready for AI.

AI thrives in environments where information flows seamlessly, where systems are unified, data is structured and workflows make sense. But in many contact centres, that’s not the case. Legacy systems still dominate, channel silos persist and business processes are often a patchwork of “temporary” fixes that became permanent years ago.

Automation and consolidation should come before AI. They are the plumbing that allows intelligence to flow. Without them, all you’re doing is adding another layer of complexity, and this one magnifies the inefficiencies.

Let’s address the human factor. Many leadership teams are seduced by the promise of AI efficiency without doing the unglamorous groundwork. It’s far easier to announce an “AI transformation initiative” than to spend months rebuilding workflows or cleaning up data.

AI can do a lot but it doesn’t forgive laziness. If your customer journeys are broken, AI will expose that faster than any human ever could. If your process design lacks empathy, AI will replicate that indifference at scale. Technology amplifies culture and, in most cases, it amplifies dysfunction just as effectively as it amplifies success. AI will not save a disjointed customer experience. It will simply make it more visible.

Before any organisation even considers AI in its customer engagement ecosystem, three conditions must be met:

1. Data structures: Your customer data must be connected, consistent and trustworthy. You can’t feed a model fragmented, conflicting inputs and expect coherence.
2. Customer journey understanding: You need to know what your customer actually experiences, not what you think they do. Map it, test it and feel it from their side.
3. AI literacy within the organisation: Leaders and frontline teams need to understand what AI can and cannot do. Without internal literacy, expectations spiral and projects stall.

If any of these are missing, stop. You are not ready.

The Sabio Group recently highlighted that 95% of enterprise AI projects fail to show measurable impact, primarily due to lack of expertise in operational integration and process design. McKinsey echoed a similar sentiment, noting that companies that achieved success were those that had first addressed process, data and tooling fundamentals.

In other words, the ones that got it right didn’t start with AI, they started with clarity. They used automation and workflow intelligence to clean up the environment first. Only then did AI become the accelerator it’s meant to be.

If you want a roadmap that actually works, it looks like this:

1. Map your end-to-end customer journey. Identify friction points, escalation loops and gaps between intent and resolution.
2. Automate your bottlenecks. Focus on repetitive tasks that waste human potential like data entry, verification, routing.
3. Consolidate systems. Reduce the cognitive load on your agents. Give them one workspace, not 10.
4. Clean your data. Make sure what you’re feeding AI is accurate, current and aligned across departments.
5. Pilot AI purposefully. Don’t “do AI” for optics. Apply it surgically where it solves a defined business or customer problem.
6. Monitor the impact holistically. Look beyond deflection rates. Measure sentiment, satisfaction and the human experience at both ends of the interaction.

AI is not the enemy of good service. Done right, it can be the most powerful tool we’ve ever had to improve efficiency, insight and experience. But rushing in without the right scaffolding doesn’t make you innovative. It makes you a little reckless, frankly.

There’s no shame in slowing down to do it properly. Process, automation and data hygiene aren’t the boring parts of transformation. They are in fact the parts that make transformation possible.

So yes, we are getting ahead of ourselves. But we don’t have to stay there.

Let’s build the systems that can actually sustain the intelligence we’re chasing. Because the real measure of progress won’t be how fast we deploy AI, it will be how intelligently we prepare for it.