Agentic AI Will Run 68% of Customer Service by 2028.

In a landmark and somewhat shocking global study released this week, Cisco revealed a very bold prediction: by 2028, 68% of customer service and support interactions will be handled by agentic AI, autonomous systems capable of acting, deciding, and learning with little to no human oversight.

I hope the headline is attention-grabbing enough, but what lies beneath it is far more significant and rather unsettling.

Let's be clear, this isn’t about chatbots getting smarter. This is about entire layers of human interaction being stripped out of the customer journey and replaced by intelligent, self-acting systems. Cisco calls it the “agentic future.” But for brands, leaders, and workers, it’s actually a big fork in the road. And not enough people seem to be asking the deeper questions.

Because here's the thing, while agentic AI might solve tickets faster, there’s a very real growing cost we’re not spending enough time on or calculating:

Trust. Jobs. Identity. Emotional presence.

If we’re not super careful right now, that 68% could become the very symbol of just how disconnected “service” becomes from the people it was meant to serve.

So what’s Fuelling the Rise of Agentic AI?

Cisco’s research was drawn from nearly 8,000 leaders across 30 countries and paints a much clearer picture of the growing complexity in enterprise environments. The IT stack has become very bloated. Customers want immediacy which has put leaders under increasing pressure to cut costs, scale faster, and reduce friction.

Enter agentic AI.

Unlike traditional automation, agentic AI isn't just programmed, it has agency, that means it thinks for itself. It can reason. It holds onto memory. It makes independent decisions. It learns from context, adapts to new information, and executes actions without waiting for a human to approve them.

93% of survey respondents believe this technology will create more personalised, proactive, and predictive customer experiences. And 88% say it will help accelerate strategic goals like digital transformation, cost efficiency, and IT resilience. 

In other words: it works. And it works well.

But here’s the problem: just because it works, doesn’t mean it’s working for everyone. 

What We’re Not Talking About: Trust, Transparency & the Human Toll

Cisco’s report briefly touches on the need for ethical implementation and robust governance, 99% of respondents said it’s important. But that statistic, while it might be comforting on the surface, raises a deeper truth: the very success of agentic AI will depend on trust,  and right now, trust is more fragile than ever before.

People don’t trust what they don’t understand. It's kind of simple. And very few customers, or employees, truly understand agentic AI. When a system makes a decision that no human directly approved, the accountability chain is broken. And this is where the trouble begins.

• Who do I blame when something goes wrong?

• Who do I speak to when I feel unheard?

• How do I trust a process I can’t see, or challenge?

  
  

The more autonomous these systems become, the less visible the humans are behind them, and the more power quickly shifts away from the people they were actually or apparently built to serve.

Our customers might get a faster resolution. But what they’re left with is a colder, more opaque interaction, one that quietly erodes all the human glue that brands rely on to build loyalty.

So Now Let’s Talk About the Jobs

The report does not explicitly address the workforce implications nor does it suggest who's going to. But we need to. Because 68% of service interactions don't automate themselves, they replace someone.

If Cisco’s forecast is right it means by 2028, most help desks, support roles, live chat agents, and service centers as we know them will be either eliminated or dramatically and radically reshaped. That’s millions of jobs globally. Not hypothetical jobs but  real people with rent to pay, kids to raise, and experience in human connection.

We can talk all we want about “upskilling” and “strategic reallocation.” And if you are like me it seems a little bit of a “go to” at times. But history has shown time and time again that technological disruption moves faster than human reintegration. If brands and governments don’t move quickly and urgently to plan for this shift, we’re not just heading toward a more efficient service industry, we’re heading toward a massive emotional and economic reckoning.

And just to make it more interesting, here's the twist. The very customers these brands hope to serve with AI? Many of them are the same people who will be displaced by it. I will try and make it clear. It’s not a future problem. It’s a today problem.

So What’s Left for Humans?

So if we are now assuming agentic AI takes the 68%. What happens to the remaining 32%?

That 32% becomes the difference between a brand you remember and a brand you almost tolerate. It is now the battleground for trust, intimacy, and long-term loyalty. And it becomes the space where humans still have the advantage, not because they’re faster, but because and I have always said, they feel.

In that 32%, brands will need to invest more, not less. These interactions must now carry the weight of what’s been lost elsewhere. They must be slower. More present. More human.

The irony is this: as AI gets more powerful, human presence becomes more premium. Not in volume, but in value.

Final Thought: Progress Without Presence Is Just Performance

There is no stopping agentic AI, the time for rolling this out correctly has long gone. Personally I don't want to either. The potential for improvement in our lives and world are real and the benefits are greatly tangible. The innovation is historic.

But we must stop pretending that automation and experience are interchangeable.

We have to focus and build systems that don’t just scale capability, but actually protect connection.

We must lead brands that don’t just optimise for convenience but they design for trust. And we must ask ourselves one hard question:

“If your customer only interacts with a machine… will they ever truly connect with your brand?”

That’s what’s at stake. Not just jobs. Not just speed. Not just efficiency. The soul of service.

And we’d better get that part right.

By Nick Gray, Founder — IGU Global