Dive Brief:
- The vast majority of business leaders — 81% — said they are using AI agents today, according to a Zendesk report released Tuesday. Zendesk surveyed more than 3,100 consumers and over 4,400 business respondents across 10 countries.
- Business leaders report using a median of 10 customer-facing AI agents and plan to use 15 next year. Most of these AI agents deployments are specialized agents built for defined workflows, such as a returns AI agent that can autonomously check an order, verify whether it qualifies under company policy, collect the required information, initiate the return and update the customer.
- Nearly 4 in 5 business leaders said they need to create formal oversight roles to build, configure, improve and supervise their AI rollouts and scale them.
Dive Insight:
The promised proliferation of AI agents in the workforce has arrived. Now companies need to figure out how to manage this digital labor — and who has the skills to do it. These skills range from building and configuring AI agents to knowing what processes to automate and when.
Companies are turning to their existing customer service staff, but it’s not as simple as just transferring them to a new department.
Three-quarters of business leaders said they are reskilling their existing service staff to fill those AI oversight roles. However, less than one-third of human agents said their organization is actively helping them develop the skills they need to oversee AI agents. Another 44% of agents said they don’t see a clear career path for people who cultivate AI management and oversight skills.
Contact center professionals already possess considerable, hands-on operational knowledge that is transferable to supervising AI-powered processes. Experienced agents have handled exceptions and complex situations, can likely recognize recurring service failures and friction points, and identify gaps between standard operating procedures and solving actual customer needs.
“Experienced contact center professionals could become some of the most valuable people in an AI-enabled service organization,” Behnam Behzadifar, a CX expert and senior strategist with Mofid Securities, told CX Dive in an email. “However, I would distinguish between simply transferring existing contact center employees into AI supervision roles and deliberately developing a new professional capability.”
Zendesk found that two-thirds of business leaders do not have a clear picture of what “AI orchestration” roles would require — which helps explain the gap between what business leaders say they are doing with respect to reskilling and what human agents are experiencing.
The surveyed organizations didn’t have a sense of what skills would be necessary, but Zendesk’s report identified five areas that CX leaders should develop. These range from designing how AI agents, human agents, workflows and data come together to solve customer issues to assessing the performance of the AI-enabled workforce through metrics like resolution quality, cost, business outcomes and customer impact.
Behzadifar envisions several possible AI supervisory roles for human agents, including AI trainers and knowledge curators, AI operations specialists, conversation quality analysts and conversation designers.
But these roles require structured reskilling, Behzadifar said. And those roles include a multidisciplinary skill set, including understanding large language model capabilities and limitations, data and knowledge engineering, and guardrail implementation. Customer centric skills like journey mapping and interaction design are also key.
“The most successful organizations will be the ones that understand where autonomy creates value, where human judgment remains essential, and how to orchestrate both to deliver better customer outcomes,” Behzadifar said. “Ultimately, I believe the objective should not be to maximize AI autonomy, but to optimize the balance between AI autonomy, human judgment and customer value.”