AI to Cut 350,000 US Contact Centre Jobs by 2031, Forrester Finds

What happened

Forrester Research forecasts that the number of human customer service agent roles in the United States will fall by approximately 350,000 between now and 2031, as artificial intelligence and automation absorb a growing share of routine contact-centre work. The projection, reported by No Jitter, represents one of the more concrete workforce estimates to emerge from a major analyst firm on the subject of AI's displacement effect in CX operations.

Despite the headline figure, Forrester's analysis stops well short of predicting the elimination of human agents altogether. The research points instead to a structural shift: AI handles high-volume, low-complexity interactions, while human agents are repositioned toward escalations, emotionally sensitive conversations and cases that require judgement beyond what current models can reliably provide. The forecast therefore describes a smaller, differently skilled agent workforce rather than an absent one.

Why it matters

For customer experience leaders and service designers, the Forrester forecast reframes the automation debate from "will AI replace agents?" to "what does the human role become?" That is a meaningfully different design problem. Organisations that treat headcount reduction as the primary return on AI investment risk hollowing out precisely the human capacity — empathy, contextual reasoning, recovery from failure — that drives loyalty in high-stakes moments. Behavioural economics research consistently shows that customers weight negative service experiences more heavily than positive ones; the interactions most likely to remain with humans are therefore the ones that carry the greatest emotional and reputational consequence.

Service designers will also need to grapple with transition risk. A contact centre that reduces agent numbers faster than it builds AI reliability creates a capability gap that customers experience directly — longer resolution times, mishandled escalations, and eroded trust. The workforce question and the experience quality question are, in practice, the same question.

By the numbers

• 350,000 — projected reduction in US customer service agent jobs by 2031, according to Forrester Research.
• 2031 — the forecast horizon over which this workforce contraction is expected to materialise.

The Renascence take

The risk most organisations will miss is not the job loss itself — it is the quiet degradation of institutional knowledge that leaves with those roles. Every experienced agent carries an informal understanding of where processes break down, what customers actually mean when they complain, and how to improvise when the system fails. That knowledge rarely lives in a knowledge base.

Cutting 350,000 agents without first extracting and encoding what they know is not an efficiency play — it is an experience liability. The behavioural principle at work is the peak-end rule: customers remember how a journey ended and its most intense moment, both of which tend to involve a human. Operators who want AI to carry more of the load should invest equally in capturing agent expertise to train those models, and in designing clear, low-friction escalation paths so that the human moments that remain are genuinely good ones — not just the moments AI could not handle.