AI Agents and the Changing Face of Customer Service Telephony

Is this the death knell for hold music?

Evolving Trends in Customer Service Telephony

With the advent of AI agents, are we are likely to see a resurgence in telephony as a preferred method for customer service interactions? I was interested to hear Gadi Shamia and the team at Replicant offer further insights into this evolving debate during their recent Form1 Partners briefing. They highlighted newly enhanced features across their product range, while providing compelling accounts of customer successes and tangible benefits observed across various use cases.

The Shift Away from Traditional Phone Support

Over the last several years both large and smaller retailers, and service providers, have actively encouraged consumers to move away from calling customer service by phone. Instead, they have steered them towards using mobile applications, websites, or basic chatbot solutions. The primary motivation behind this strategy? To reduce the significant costs associated with employing human agents — even when such roles are outsourced to countries with lower labour costs.

And the results of this shift? Mixed.

Some retailers have successfully streamlined certain processes—such as item 'Returns'—allowing customers to complete these tasks with a few clicks and immediately receiving a QR code or mailing slip for printing. However, more complex or less common queries, scenarios where the standard process has not worked, or situations where a customer has multiple questions, can lead to frustration and lengthy interactions. For the consumer this has resulted in much wasted time and a diminished perception of the retailer's service quality and ultimately brand reputation.

Rolling out the Agentic AI Contact Centres

With the advent of genuine Agentic AI contact centre platforms, like Replicant, the landscape is beginning to change. These AI agents are capable of engaging in natural language conversations, understanding customer intent, and interacting empathetically while adhering to the retailer’s policies and standards.

Unlike human agents, AI agents do not suffer from fatigue, are available on weekends and holidays, and can scale instantly to meet fluctuating demand with a virtually limitless number of agents.

As people grow increasingly accustomed to interacting naturally with technology—evidenced by the widespread use of products such as Amazon Alexa—the preference for speaking rather than typing is becoming more pronounced, especially when it avoids long waiting times, poor quality hold music, and delays caused by slow systems. If AI agents can efficiently and accurately resolve queries, most consumers would likely choose to speak rather than type.

Ongoing Need for Human Oversight

Despite these advances, the importance of human supervision and escalation pathways remains. It is crucial to ensure that AI agents are well-trained on the most common use cases and can continue to learn autonomously. When properly implemented, these systems have the potential to resolve over 80% of customer calls without the need for human intervention. The role of human customer service agent is redirected to managing the more complex 20% of calls and, with more time to focus on issue resolution rather than wading through call volumes, is likely to become a more respected 'can do' specialist...and could see these roles coming back on-shore.

Reap the Advantages

As with all query based interactions there are further benefits to broadening the implementation of Agentic AI. Each call will generate valuable data points. Analyse these and you not only identify recurring issues that enable improvements in service delivery — but these insights present opportunities such as targeted follow-ups or up-selling, either during or after the interaction.

What are your thoughts? Looking to explore more, get in touch david@eleos.london