AI Is Not Cutting Service Jobs. It Is Making Them Harder.

The expectation was mass redundancy. Gartner surveyed 321 customer service and support leaders worldwide between September and October 2025, and found that 31% had implemented or were planning frontline workforce reductions through layoffs in response to AI through the first quarter of 2027.

In the same survey, 85% were adding new tasks and responsibilities to frontline agent roles.

That is the finding worth sitting with. AI is not emptying the service floor. It is rewriting what the people on it do, and it is doing so at the exact moment those people became the most important part of the customer experience.

What is actually happening to the role

The Gartner data describes redesign rather than elimination. Eighty percent of service and support leaders reported pressure to make workforce changes as AI reduces contact volumes and improves agent efficiency. Sixty-three percent are reducing frontline headcount gradually through attrition while reallocating agent capacity toward higher-value responsibilities. Seventy-five percent are shifting agents into entirely new roles inside the service organisation.

Kathy Ross, a Vice President Analyst in Gartner’s Customer Service and Support practice, framed the risk for leaders as needing a plan for how they reshape the workforce, “otherwise a plan will be handed to them”.

So the frontline job in 2026 involves more tasks, broader responsibility and, for three quarters of organisations, an outright change of role. Gartner describes the scope of the role expanding; the judgement that this makes the job harder rather than merely different is mine, and the rest of this piece argues for it.

Why this makes employee experience a CX constraint

Here is the mechanism, and it follows directly from how AI deflection works.

AI absorbs the simple, repetitive, quickly resolved contacts. That is what it is good at and what it is bought for. Whatever reaches a human is therefore, by design, the residue: the complex cases, the exceptions, the angry escalations, the problems the automation could not close. The easy wins that used to break up an agent’s day are gone, because those were the first things automated.

Now put that next to what customers demand. In a separate Gartner survey of 3,566 B2B and B2C customers conducted in February and March 2026, 87% said access to a human agent is essential when companies use generative AI, and only 27% would be willing to try a chatbot again after a negative experience. Both figures come from that one survey, reported in two separate Gartner releases.

Those two facts interlock. The human route is the thing customers insist on, and the thing that salvages a failed automated attempt you will not get a second chance at. It is staffed by people whose caseload has been concentrated into the hardest work and whose job description is being rewritten underneath them.

If employee experience degrades under those conditions, the customer experience strategy loses the one component it cannot afford to lose. That is not a morale argument. It is a design dependency.

What the evidence says about the link

The relationship between employee engagement and customer outcomes is one of the better-evidenced findings in management research, and it is worth citing precisely rather than loosely.

Gallup’s Q12 meta-analysis accumulated 736 research studies across 347 organisations in 53 industries, covering 183,806 business and work units and 3,354,784 employees in 90 countries. It applied Hunter-Schmidt meta-analytic methods and states that it includes all available Gallup studies, published or unpublished, so it carries no publication bias risk.

The true score correlation between employee engagement and composite performance is 0.49. Business units in the top half on engagement nearly double their odds of above-average composite performance within their own organisation, and raise those odds by 2.33 times when the comparison is drawn across every organisation in the database.

On customer outcomes specifically, the median difference between top-quartile and bottom-quartile units was 10% in customer loyalty and engagement. The same comparison showed 23% in profitability and 14% in productivity measured through production records and evaluations.

Two caveats, because this research gets overstated constantly. These are quartile comparisons and correlations, not proof that engagement causes loyalty; a well-run unit plausibly produces both. And 10% is a median across a very large and varied database, not a return anyone should put in a business case for a specific programme.

What the data does establish is that the association is consistent across organisations, industries and countries, which is a stronger claim than most people realise and a weaker one than most people make.

Customers still prefer the human for the decisions that matter

There is a related Gartner finding that cuts against the idea of the agent as a legacy cost. In a survey of 5,801 US customers conducted in January and February 2025, 54% said they trust human agents more than AI for product or service recommendations, against 32% who trust AI more.

Recommendation is advisory work. It is the high-value, judgment-dependent interaction that organisations say they want agents redeployed into, and customers are already telling them the human has the advantage there. The 75% shifting agents into new roles are, on this evidence, moving them toward work customers actively want them doing.

That only pays off if the people make the transition successfully, which is an adoption and capability problem rather than an org chart problem.

What this means in practice

Treat the role change as a transformation programme, not an announcement. Moving three quarters of your agents into new roles is a change and capability exercise with training, new competency models and a realistic ramp. Organisations that reassign the role and leave the person to work it out produce exactly the attrition they were hoping to avoid.

Rebalance the caseload deliberately. If automation takes every easy contact, the human queue becomes uniformly difficult. That is a predictable path to burnout and it is a choice, not a consequence. Deliberately routing some straightforward volume to humans costs efficiency and buys sustainability.

Measure the employee side of the journey with the same rigour as the customer side. Most organisations can quote their CSAT and could not tell you agent effort, time-to-competence in the new role, or attrition among people whose job changed in the last year. If it is not measured it is not managed, and the EX metrics are usually the missing half.

Give agents the AI, not just the customers. Agent-facing AI that drafts responses, retrieves context and summarises a case history reduces the effort of handling the hard cases that automation now routes to them. The investment case is better here than in the customer-facing bot, and it is consistently funded second.

Involve the frontline in designing the journey. Agents know which steps generate repeat contacts and which policies cause the escalations, because they absorb the consequences daily. Journey work that excludes them is mapping the process rather than the experience.

What this data does not tell you

The Gartner workforce survey covers 321 leaders globally with no GCC breakdown, and its fieldwork was September to October 2025, so it describes intentions formed more than a year ago. The 5,801-customer trust survey is US-only and was conducted in early 2025. The Gallup meta-analysis is global and includes Middle East work units, but it reports cross-organisational medians rather than anything specific to this region or to contact centres.

I have not found comparable GCC-specific data on service workforce redesign. Anyone reasoning about the Gulf market is extrapolating, including me, and should say so.

The connection most organisations miss

Customer experience and employee experience get owned by different functions, measured on different cycles and funded from different budgets. In a service operation being reshaped by AI, that separation stops being an organisational quirk and becomes the thing that breaks the strategy.

The automation is already deployed. The customers have already told us they want a human available when it fails. The only open question is whether the humans who remain are equipped, staffed and supported well enough to be worth reaching, and that question is answered in the employee experience or not at all.

Martin Alva

Marketing, Brand & Digital Transformation Leader, United Arab Emirates

Senior Manager, Employee Experience (Brand Positioning, Digital Platforms & AI) at Space42 (A G42 & Mubadala Company). 15+ years in marketing, brand and digital transformation across MENA and Europe. 5 MENA Effie Awards.

More about Martin Alva → Connect on LinkedIn →