How Do New Contact Center Agents Learn When AI Handles the Easy Contacts?

In many contact centers, new agents now learn on harder work from their first week, because AI is absorbing the simple contacts that used to serve as their training ground. For decades, a new hire built confidence on balance checks and order status questions before moving on to complaints and exceptions. As those transactional contacts move to self-service and AI agents, that gentle on-ramp disappears, and new agents meet the difficult conversations far earlier in their tenure.

The potential scale of the shift is large. McKinsey's analysis of millions of customer interactions across more than 30 organizations found that 50 to 60 percent of interactions remain transactional, and that is the kind of volume AI is increasingly able to handle. As that share moves away from human agents, time to proficiency becomes a capacity variable that workforce leaders need to plan for and protect, alongside the forecast and the schedule.

Why does AI make it harder for new agents to learn?

AI makes it harder for new agents to learn because the transactional contacts it handles first were also the low-stakes repetitions where new agents built product knowledge and confidence. Every simple contact taught a new hire something about the systems and the rhythm of a customer conversation, at a moment when a mistake carried little cost. Remove a large share of those contacts and new agents get far fewer chances to practice before the stakes rise.

What remains for human agents is the harder work. When AI resolves the routine requests, the contacts that reach people are disproportionately the exceptions and emotionally charged conversations that need judgment. A new hire who would once have handled dozens of easy contacts in a shift now spends that shift on work that a few years ago would have been routed to a tenured agent. That is a heavier load to carry while still learning the systems, and it raises the stakes on how the first months of tenure are designed.

Can AI help new agents get up to speed faster?

Yes, and the strongest research on the question shows AI assistance compressing the learning curve for new agents in particular. A study by Erik Brynjolfsson, Danielle Li, and Lindsey Raymond, published in The Quarterly Journal of Economics in 2025, used data from 5,172 customer support agents at a Fortune 500 software firm during the staged rollout of a generative AI assistant. Productivity, measured as issues resolved per hour, rose 15 percent on average and roughly 30 percent for less-experienced and lower-skilled workers, with minimal impact on the most experienced and skilled agents.

The tenure finding is the one workforce planners should notice. According to the study, agents with two months of tenure who used the AI tool performed as well as agents with more than six months of tenure working without it. The study also found that customers expressed more positive sentiment in conversations with agents who had the assistant, and that access to it was associated with higher agent retention. The researchers suggest the AI may have spread the practices of the most capable agents to newer ones, giving them access to experience they had not yet built.

So AI removes the traditional training ground and offers a faster path at the same time. Which effect wins in a given operation depends on how deliberately onboarding is designed around both.

How should onboarding change for a contact center where AI handles routine work?

Onboarding works best when it builds deliberate practice into the program, so new agents rehearse difficult contacts in a safe setting before they meet them live. Simulation is the most direct way to do that. McKinsey, discussing simulation-led onboarding and training, reports seeing 20 to 30 percent reductions in agent time to proficiency in its client engagements, with simulated complex scenarios letting agents practice before they meet them live.

Nesting deserves a redesign as well. Nesting is the period after classroom training when new agents take live contacts with extra support, and in a hybrid operation it works best with graduated complexity. Routing can deliberately give nesting agents a controlled share of complex contacts that grows week by week, with AI assistance available throughout and a supervisor or mentor close at hand. The goal is to recreate the progression from easier to harder work that the old contact mix used to provide naturally.

AI assistance belongs inside this design from the first day of nesting. The gains in the Quarterly Journal of Economics study came from an assistant that suggested responses during live conversations, which meant new agents saw what a strong answer looked like at the moment they needed it. Supervisors can deepen that learning by reviewing a sample of assisted conversations with each new agent and discussing why the suggested approach worked, so the knowledge stays with the agent on the contacts where the assistant has less to offer.

Coaching closes the loop. Quality evaluations for new agents are most useful when they focus on the handful of behaviors that matter most for the complex contacts those agents now handle, and when coaching follows quickly enough to shape the next shift. A new agent who gets targeted feedback within a day is likely to learn faster than one who waits for a monthly review.

How do you plan capacity around new-agent ramp time?

You plan capacity around ramp time by modeling new hires at partial productivity that rises week by week, and by scheduling training and coaching as protected time in the plan. A capacity plan that counts every new hire as a full agent from their first day on the floor will understate the staff needed in any month with a large hiring class. Forecasting handle time by tenure cohort, along with how much of the contact mix each cohort can take, gives a more honest picture of how much work the floor can actually absorb.

Training and coaching time belongs in the schedule as planned shrinkage, and it needs protecting during the day. When an interval runs hot, coaching sessions are often the first thing canceled, and in an operation where new agents depend on coaching to reach proficiency, those cancellations carry a cost that shows up weeks later. Measuring time to proficiency by hiring cohort, and comparing cohorts as onboarding changes, tells you whether the investment is working.

This is where Aspect connects planning with performance. Aspect Workforce builds training and coaching time into schedules as part of the plan, and Aspect Performance gives supervisors the scorecards and coaching workflows to track how each new agent is progressing. Aspect Intelligence flags intraday risk early, so planners have room to rebalance before a busy interval forces them to cancel the coaching that new agents need most. The planner stays in control of those trade-offs, with the system surfacing them while there is still time to choose.

What this means for your workforce strategy

As AI handles more of the routine work, the way people become good at the job has to be designed on purpose. The contact centers that adapt will treat time to proficiency as a planning input and invest in onboarding that uses simulation and AI assistance to speed up the learning the old contact mix used to provide.

Human work in the contact center is getting harder, and the agents doing it deserve an on-ramp built for the work they will actually face. The operations that build that on-ramp give their newest agents a realistic chance to succeed at the complex contacts that now define the customer experience.

‍