How to test for customer service roles today

There are roughly 2.7 million customer service representative jobs in the United States, with about 289,500 openings per year. That sounds like a massive, stable workforce. It isn't. The BLS projects a 5% decline over the next decade, driven by the same force reshaping every support operation in the country: automation is absorbing the work that used to keep most of those seats filled.

We talked to Ken McMahon, Chief Customer Officer at Nextiva, about how this shift is changing what companies should look for when they hire customer service agents, and why most hiring processes haven't caught up.

The customer service job has changed, but most hiring tests haven't

AI spending in customer service jumped 38% year-over-year in 2025, according to a Gartner survey of service and support leaders. Technology budgets are up 15% overall, and AI initiatives now account for 13% of total service budgets, a number that's growing at the expense of labor and overhead.

The result is a workforce that's getting leaner and a job description that's getting more demanding. 

When AI handles password resets, order tracking, return initiation, and FAQs, the tickets that reach a human agent are often complex:

• Multi-system issues that require navigating three platforms simultaneously. 
• Compliance-sensitive conversations where a wrong answer creates legal exposure. 
• Customers who are already frustrated because the bot couldn't help them.

"Two years ago, a new customer service hire could get by on friendliness and basic product knowledge," McMahon says. "But the interactions reaching agents today are more complex, and emotionally charged, and they’re often higher stakes from a customer experience perspective. If someone can't think on their feet and navigate ambiguity, they're going to struggle."

This creates an obvious problem for hiring. Most customer service assessment tests were designed for the old version of the job. They focus on typing skills and an ability to follow a script. And while those skills do still matter, they’re no longer enough to assess whether or not your agent will be able to handle complex, compliance-focused, or emotionally-charged conversations.

What a customer service skills test should measure now

O*NET's profile for customer service representatives lists the core competencies: active listening, speaking, reading comprehension, critical thinking, service orientation, complex problem solving, time management, social perceptiveness. Those haven't changed. What's changed is how much weight the harder skills now carry relative to the easier ones.

Here's what actually separates a strong hire from a warm body in the current environment:

Judgment under pressure. Can the candidate make a good call when the playbook doesn't cover the situation? AI escalates the weird stuff. Agents need to handle it without freezing, guessing, or defaulting to "let me transfer you to a supervisor."

De-escalation and emotional intelligence. Nearly 90% of consumers say fast, accurate resolutions determine whether they'll buy from a brand again. But speed without empathy backfires. Three in five consumers say agents are only half-listening when they call. The agents who retain customers are the ones who make people feel heard while actually solving the problem.

System fluency. The average support agent in 2026 toggles between a helpdesk, a CRM, an order management system, and sometimes a knowledge base, shipping platform, or compliance tool. The ability to move between systems quickly while maintaining context isn't a nice-to-have. It's a core job function.

Adaptability. The tools are changing fast. The AI capabilities your agents work alongside this quarter might look totally different next quarter. Agents who can absorb new workflows, pivot when a process changes mid-shift, and teach themselves a new platform without needing to be retrained from scratch are the ones who survive in a role that's being reshaped in real time.

"Adaptability is the one most people underweight," McMahon says. "You can't teach someone to be comfortable with major changes every few months. That's either in the person or it isn't."

How customer service skill testing can vary by industry

A customer service agent at a retail company handles returns, exchanges, and shipping complaints.

 An agent at a healthcare organization handles insurance questions, appointment scheduling, and conversations where HIPAA compliance isn't optional. 

And an agent in financial services handles account disputes, fraud alerts, and transactions where a compliance misstep can trigger regulatory consequences.

These are fundamentally different jobs that happen to share the same title. Testing them with the same generic assessment is like using the same interview questions for a pediatric nurse and an ER surgeon because they both work in healthcare.

"Industry context matters more than people think," McMahon says. "An agent who thrived in ecommerce support might completely flounder in healthcare, not because they're less skilled, but because the domain knowledge and the stakes are different. When you're hiring for regulated industries, you need to know before the first day whether someone understands what they can and can't say."

Here’s how priority skills matter by industry: 

Why you still need customer service agents in the age of AI

AI-powered customer service agents from reputable providers can be incredibly helpful as a first line of defense for customer assistance. They handle password resets, order tracking, shipping updates, and FAQ-level questions faster than any human can, and they do it around the clock without staffing constraints. About half of consumers say gen AI actually makes support interactions easier, and 96% of ecommerce companies are already using AI in some capacity.

But the companies that went all-in on replacing human agents are learning that the math doesn't hold. Gartner predicts that 30% of employees laid off in AI-driven cuts will need to be rehired by 2029, often at a higher cost than retaining them would have been. 

For some companies, this has already happened. Klarna is the most public example. They claimed its AI could replace 700 customer service reps, then quietly started rehiring. 

Nearly 90% of customers say companies using generative AI must also provide access to a human agent. They don't want AI gone, but they want both. When the bot can't fix their issue, they want a real person who can.

"The smartest thing companies can do right now is stop thinking about this as AI versus people," McMahon says. "AI should handle the transactional requests. Humans should handle the abstract, relationship matters. But you need to hire people who can actually handle that complexity, which means testing for it before they start."

And as AI absorbs the routine work and your support team is leaner because of it, every human hire carries more weight. The margin for a bad hire just got a lot thinner.

How to build a customer service test that works

A customer service skills test built for 2026 should cover three layers:

Foundational skills

Typing speed, reading comprehension, basic math, communication clarity. These are still prerequisites, and they're easy to measure. A candidate who can't write a clear email or process a simple calculation shouldn't move forward regardless of how strong their soft skills are.

Role-specific technical knowledge

This is where industry-specific assessments come in. eSkill's assessment library includes dedicated tests for customer service in healthcare, financial services, retail, and hospitality, each built around the domain knowledge and compliance requirements that matter for those roles. 

You can also combine subjects — pairing a general customer service assessment with CRM proficiency, data entry accuracy, or written communication, for example — to match what the specific position actually demands.

Judgment and soft skills

Scenario-based questions that simulate the ambiguous, emotionally charged interactions agents now face daily. Can the candidate prioritize when three things need attention simultaneously? Do they escalate appropriately or try to handle everything themselves? Can they maintain composure and brand voice under pressure?

The goal here is to surface the ones who have the baseline competencies plus the judgment and adaptability that the role now requires, before you invest interview time in someone who looks great on a resume but can't perform under the conditions the job actually presents.

The future of customer service work

The customer service workforce is contracting, but the remaining roles are more demanding, more specialized, and more consequential than they've ever been. A customer service test designed for the old version of the job — one that measures typing speed and script adherence — won't identify the agents who can thrive alongside AI in a leaner support operation.

Test for what the role actually requires now: judgment, emotional intelligence, industry-specific knowledge, and the ability to handle the interactions that AI can't.