Across the Universe
Wouldn't it be ideal if consumers could call a single number at a company for customer support and get their problem fixed by any agent that picked up the phone, regardless of what product they own? In their endless efforts to strike the optimal balance between reducing costs and improving customer service, contact center managers have begun to examine this idea of the universal agent, agents theoretically capable of handling any contact, via any channel.
This notion certainly grabs the attention of contact center managers, the poor folks who have been repeatedly asked to hold the line on costs. In fact, universal agents sound like a panacea for cost control, as such a setup would require both fewer agents and a less complicated technology infrastructure.
The arguments in favor of this universal agent model often include improvement in customer service and customer experience, as well. Fewer transfers and escalations of contacts should mean that first-call resolution rates will rise--and better first-call resolution rates translate into higher customer satisfaction.
Like most supposed panaceas, however, this model raises at least as many problems as it might solve. To succeed with a universal agent model, every
agent must have a sturdy grasp on all
of the possible types of transactions that could occur in that customer interaction organization. Agent skill sets, and the methods for quantifying and tracking those skills, developed for a reason. At the most basic level, in a multicultural and multilingual society, it makes little sense to only hire agents completely fluent in all of the major languages that might be used in the contact center, especially for large companies with hundreds or thousands of agents.
A truly universal agent would require a lot of behind-the-scenes technology. Much of that technology would actually improve the agent and customer experience regardless of whether the agents are specialized or universal. A single sign-on and a unified desktop would reduce the frustration and resource consumption of repetitive data entry; agents would also be spared having to switch between numerous applications. The simplified desktop should be an ideal toward which organizations consistently strive, but the universal model basically requires
such a desktop.
Training issues would also be exponentially more complex for universal agents. Customer-facing employees would need to both understand every transaction possible with a customer in that contact center and
know how to operate the tools and applications required for those processes. To add complexity atop complexity, the universal agents would need to have equal facility providing that support over every channel, including chat, email, and phone, as well as possibly collaborative applications.
Still, a universal agent model can
make sense--but only under very limited circumstances. For example, universal agents could be a fit for some B2C companies that have a single product line--even if that line is a high-volume one. Those agents would have only a limited set of skills to master; customer concerns would center on just a handful of topics. This model even applies to multilingual centers with simple processes, such as directory assistance, where all agents must be fluently bilingual. But as companies grow--and that does seem to be the idea for most companies--the number of products, product areas, and even business units expands, undermining the aptness of the universal agent model.
In cases where the customer-support organization must deal with such different types of support (i.e., numerous product types, product lines, etc.), organizations should focus on creating multiple skill sets, workflow rules to handle issues such as escalation, and a well-built routing engine to ensure that the best agent gets sent each contact--in short, a universal contact queue with a routing-engine traffic cop to send contacts to empowered and well-trained agents.
Ian Jacobs is a senior analyst in Frost & Sullivan's contact center practice. Contact him at email@example.com.