As AI handles more routine customer interactions, the contact center workforce is contracting — and the human agents left over must handle the more complex calls.
- As bots take over routine customer service interactions, companies need fewer frontline agents, while the humans who stay are increasingly handling high-stakes issues.
- The “agent of the future” needs a broader skillset. Empathy still matters, but critical thinking and the ability to navigate complicated interactions are becoming essential — meaning agents may also require more training and take longer to reach proficiency.
- Fewer agents could mean better careers — if companies reinvest their AI savings. Companies like Lyft are using AI-driven savings to upskill agents and create clearer career paths. For contact centers, the new hiring challenge is finding people with the right skills who will also stay long enough to justify that greater investment.
According to some sources, AI is not actually taking all the jobs, although it’s a convenient excuse when it comes to layoffs. That is, except when it comes to customer service.
In customer service overall, Forrester research estimates that half of roles will be lost to AI by 2030. The Bureau of Labor Statistics also attributes a decline in the number of customer service roles — including contact center positions — to automation and technology advancements.
Conversational LLM agents have made it possible for both BPOs and in-house contact centers to relegate the easier calls to AI. For corporations, that means millions of dollars saved in workforce costs. For the real human agents left over, it means having to take harder calls and handle more complex work. And what does that mean for training times and the cost of attrition in turn? It’s a chain of cause-and-effect many contact centers must plan for now.
We’ll begin with the first link in the chain — the fact that the number of customer service agents that contact centers employ is objectively, truly lowering.
The customer service workforce is contracting
As AI takes a larger role in the contact center, organizations are either firing agents or, because of high turnover in the industry, waiting for natural attrition to decrease the size of their workforce. Forrester’s forecast — that half of customer service jobs will be lost to AI — is already in motion.
According to data from Indeed Hiring lab published in the U.S. Federal Reserve Economic Data, U.S. customer service job postings have dropped about 10% below pre-pandemic levels. (Overall job postings, in comparison, are still above their pre-pandemic levels.)
As customer service inquiry volume is untangled from the actual size of the labor force, contact centers are able to handle more customer inquiries while shrinking the number of human agents on the line. This trend is spotlighted in a July 2026 article from the LA Times, which notes that contact centers in service to companies like Microsoft, Uber, and Hyatt are experiencing broad layoffs as the organizations pivot to automation. As a result, the companies report savings in the hundreds of millions of dollars — Microsoft, for example, estimates annual savings of $750 million.
Will cuts stick? Customers still often want to speak to a human — and they’ll ask their AI agent until it budges and connects them to a real agent. “Zeroing-out” — when a customer presses the zero key on their phone keypad or repeatedly says "agent" to an IVR system to bypass automation and reach a human — has always been a common occurrence in customer service, but now it goes double for the AI era. Gartner predicts half of companies that laid off customer service workers due to AI will rehire them by 2027, only under new titles.
For the contact center agents that do stay, the type of work they do is changing. And in many cases, it’s becoming more complex.
The work of contact center agents is changing
Contact centers are also pursuing workforce redesign over (or in addition to) shrinking their workforce. A Gartner survey found that 85% of service and support leaders are expanding human agent responsibilities as AI reduces contact volumes, while 78% are shifting agents into different roles within service and support.
With AI taking the low-hanging calls of customer service, human agents are most obviously left taking the more complicated calls, often back-to-back — raising questions of employee burnout as agents spend more time on complex exceptions or emotionally sensitive calls. However, there is also evidence that agents prefer doing meaningful work — which includes tasks that involve variety and complexity. Others may be managing gaggles of AI agents, reviewing failed bot interactions, improving knowledge bases, and helping train or supervise automated systems.
Compare that to manual and repetitive tasks — which, on the whole, agents dislike doing. The good news is that AI can assist with these exact administrative tasks: writing post-call summaries, searching knowledge bases, suggesting actions, and surfacing customer history. As a result, agents can spend more time actually listening to and helping customers.
Agents don’t necessarily miss the rote work that AI is taking away. But there appears to be a limit: when automation removes nearly all of the easy interactions, the resulting stream of back-to-back complex or emotionally charged calls can increase cognitive load and burnout. Contact center leaders must balance workloads to include opportunities to develop skills and specialize, while also giving agents a reprieve from back-to-back calls that are cognitively demanding.
Talk to sales to learn more about how the eSkill platform can help you hire.
The silver lining
With a smaller labor force and millions of dollars saved with AI, contact centers may be primed to confront some of the biggest reasons contact center employees leave: lack of opportunity and low pay.
A Forrester report finds that while frontline agents’ pay has stagnated, Forrester expects that AI will create new customer service management jobs, for wrangling both AI agents and managing human agents alongside them — and customer service pay may shift as a result.
Consider Lyft’s approach. Lyft says its Claude-powered customer-support system cut resolution time by more than 87% and saved the company millions of dollars, which it explicitly says it reinvested in upskilling customer-support agents.
Those agents now concentrate on complicated disputes and empathy-heavy conversations, in addition to high-touch programs such as Lyft Silver — a service for older riders that includes dedicated human support, including a direct phone line — rather than juggling several routine chats at once.
Lyft says support agents now report seeing a genuine career path in the organization.
If other companies take the same approach, they could still capitalize on AI savings while still helping agents who stay find a career path in contact center work. Addressing low pay and expanding career opportunities could directly combat attrition — one of the biggest recurring (and preventable) problems for contact centers since their inception.
The old math
Contact centers and especially BPOs still live and die on a small set of operational metrics: training throughput, break‑even, speed to proficiency, and attrition. If those move in the wrong direction, margins evaporate.
In hiring for contact centers, the metric that really matters is whether new hires get through training and stick long enough to pay back the cost of hiring and onboarding.
eSkill consulted with a veteran BPO leader to get her perspective on the contact center workforce. She confirms the metrics: “The only thing that really matters is, is that person going to graduate from training, and are they going to make it to six months or whatever it is for this program and hit proficiency.”
But the old math is changing.
With the contact center workforce contracting and with training times lengthening, recruiters for contact centers need to be more intentional with who they hire — making sure new hires stick around longer and don’t fall victim to the biggest risk factors for attrition: wrong fit for the role, low pay, or lack of career development. That will take a serious overhaul for contact centers willing to put in the work.
In addition, customer satisfaction and turnover have an inverse relationship in the contact center world: contact centers with the highest satisfaction often have a turnover rate of under 10 percent, far below the industry norm, according to SQM Group. And as calls get more complex, that gap may grow: more experienced agents with more time under their belts will deliver exponentially better customer service experiences.
Agent tenure averages just 13 to 15 months. Attrition per year averages around 30 to 45 percent. Contact centers will need to improve both of these metrics to justify the cost of more in-depth training and more selective contact agent hiring.
Assessing for the new agent
84% of service leaders are revising the skills or experience needed for new agent hires, according to Gartner.
Our consultant explained: “AI and self‑service are handling simpler transactions, so the cognitive load and complexity of the interactions is going to rise, and you need a whole new profile of an agent… before we would hire for customer centricity, empathy. Well, now you need to be smart and a critical thinker because you're not going to be handling the ‘what is my bill payment due’ [question].”
As such, contact centers may need to hire fewer workers — but those they do hire will have to undergo more extensive training, need a greater range of skills, and will need to stick around longer to justify the cost of training and hiring. Hiring needs to solve for both old issues — attrition, through-put, low pay — and new ones: an expanded skillset and more intentional hiring.
The agent of the future is more knowledgeable, and needs to stick around longer. That means contact centers need to put in more effort in finding who to hire and how to keep them around for the long haul.
