Executives around the world are asking the same question: Will artificial intelligence replace agents in contact centers? While there are concerns about this being the underlying reason contact centers apply AI, few challenge its ability to increase productivity and improve the customer and employee experience (CX/EX).
Typical characteristics of customer service contact centers include the following:
- Agents undergo two to four weeks or more of new-hire training and are expected to remember the information so they can quickly respond to inquiries or perform transactions.
- Agents receive a high volume of nonstop inquiries across a variety of topics.
- Agents are expected to perform the proper level of verification for each inquiry type while not overburdening the customer.
- Agents need to navigate several operating and customer tracking solutions as they respond to customers.
- Agents are expected to handle many types of voice and digital inquiries and easily transition between channels.
- Agents are expected to handle multiple digital interactions simultaneously.
- Agents are expected to be empathetic and patient, even when customers are angry or rude.
- Agents are asked to sit and not leave their workstations for hours at a time without taking breaks.
- Agents are expected to deliver outstanding customer experience sand enhance the brand, sometimes in less than 120 seconds.
- Agents are asked to upsell within the same average handle time expectations.
Clearly, being an agent is a challenging job; these employees are expected to deliver an outstanding CX in a tight timeframe to all customers, regardless of how the employee is treated, and they are often one of the lowest-paid employee groups in the company. Here are some ways AI, especially generative AI, can help contact centers balance the automation with applications that augment agent performance, make it easier to do the job, improve productivity, and enhance the CX:
- Like employees, AI bots need to be trained and kept up to date; unlike employees, AI doesn't forget what it is taught until the information is removed from its database, knowledge base, or large language model (LLM). And even better, when the proper guardrails and controls are in place, it cannot go off-script, as a live agent can.
- AI bots can scale easily and rapidly, particularly when they are cloud-based solutions. While there is a price to scaling a bot, it is a fraction of the cost of hiring and onboarding live agents. Bots can be configured to handle a large number of inquiries or interactions without breaks or vacations, and they can move from one topic to another, as each item is a unique request. And, when an AI-enabled bot doesn't know the answer, it can keep track of the item and notify a human administrator while transferring or forwarding the request to a live agent.
- Bots can be trained and given permission to access whatever systems they need to handle customer inquiries or perform transactions. And, while there are many factors to take into consideration, bots can be more secure than live agents who can be bought. (A large volume of fraud is internal.)
- Real-time guidance (RTG) applications use conversation analytics and genAI to understand live interactions as they occur, and they use this information to deliver actionable alerts or next-best-action recommendations to agents and/or supervisors. These applications find and deliver information from disparate systems, including knowledge bases, and combine it with contextual data about the interaction to present the most appropriate information to agents and augment their performance in real time.
- Automated post-interaction summarization is a recent application that applies genAI to conversation transcripts and creates an interaction summary. The draft is typically presented to the agent for review and modification, or it can be sent directly to the CRM or servicing system. These capabilities deliver quantifiable benefits to organizations by reducing agent wrap-up time by at least 50 percent, and they lessen the pressure on agents to compose comprehensive summaries in tight time frames.
- More advanced post-interaction auto-summarization capabilities are emerging that identify follow-up activities promised during the conversation to ensure employees are reminded of their customer commitments. Additional innovations for this feature go a step further and can initiate automated workflows to complete some of the actions on employees' behalf.
AI is here to stay, and it should be used in contact centers to automate what can be automated. At the same time, it should make agent jobs better by using many of the same technologies to augment their roles.
While there are many types of AI, genAI is tailor-made for contact centers due to its inherent ability to respond to questions and other text-based inputs with well-developed answers.
As with any system, security, controls, and guardrails need to be put in place to ensure the solution stays in scope.
Contact centers should use AI to deflect inquiries that do not require complex human reasoning and to support their agents by providing them with real-time assistance as they perform a very challenging job. Over time, AI will replace some agents by performing actions that can be done more quickly and accurately by a bot. In doing so, it will also position companies to employ a higher level of agent to build their brand by delivering a truly unique and hyper-personalized experience. It's a win for customers, agents, and the company.
Donna Fluss, founder and president of DMG Consulting, provides a unique and unparalleled understanding of the people, processes, and technology that drive the strategic direction of the dynamic and rapidly transforming contact center and back-office markets. Fluss can be reached at donna.fluss@dmgconsult.com.