Advanced Intelligent Automation (IA) virtual agents shift the burden of repetitive, rules-based queries and work away from human agents. This should not be seen as a threat to South Africa’s call centres and business process outsourcing (BPO) sector, but rather as today’s “business as usual”. They will free up human agents’ time to focus on higher value conversations that impact the customer relationship and brand loyalty, something South Africans are particularly good at.
This is according to Cape BPO Advisory Board members Ryan Falkenberg, co-CEO of CLEVVA, and Paul St Quintin, Digital and Emerging Technology Lead at EY Africa, speaking at a recent Cape BPO Knowledge Sharing Session.
Automation can help BPOs
Commenting on industry concerns about job losses due to more advanced IA in the contact centre domain, St Quintin says: “We should not feel threatened, because although there is pressure to do more with what we have, the technologies today allow exactly that, with the added advantage of being able to customise the conversation using the preferred approach and channel for each of our customers.
“For the BPO sector specifically, with its traditional metrics of how many calls we handle and how quickly we get through the calls, the focus is now about how we ensure call successful closure and exceeded client experience, no matter how long it takes,” St Quintin says. “People are the core ingredient to ensuring both scenarios described above can be successfully completed.”
SA’s competitive advantage
Falkenberg notes: “South African contact centre agents have exceptional EQ - emotional quotient or intelligence. This allows them to really connect with customers. Their conversations have the ability to significantly influence the relationship customers have with a brand. They can not only build but also save relationships.
“As a result, the value of these conversations is far higher than those that simply process a specific customer request or resolve a particular query. By getting virtual agents to handle more of the high-volume informational and operational calls, human agents can be freed up to handle more of the low-volume, high-value relationship-based calls. For the South African BPO sector, this offers an exciting opportunity to differentiate globally”.
Outlining the evolution of virtual agents, Falkenberg says the first iterations – chatbots – have been ‘underwhelming’ in terms of customer experience. “The initial challenge was to get natural language understanding working. Chatbots struggled to understand what customers were asking and kept offering the wrong answers. It drove people nuts, so much so that we almost wrote off the viability of chatbots.”
Generative AI is making chatbots smarter.
Significant progress has since been made, and with the emergence of generative AI, chatbots can now offer responses that feel almost human-like. However, Falkenberg notes that while the use of large language models like ChatGPT have significantly improved the quality of chatbot responses, it has not made them better at having meaningful conversations.
“Chatbots are still very much question and answer machines (QAM). You make a request and it gives you a response. Then it waits for your next instruction. And while the answers or actions that the latest chatbots can provide are impressive, they don’t perform like human agents,” he says.
Human agents can act as both assistants and advisors to customers. At times they are asked to simply provide the right answer or perform the required action. But other times they are required to advise the customer, understand their situation, unpack their needs or identify the root cause of their problem before offering a relevant solution. This requires the ability to ask the right questions and to work out the right answers. Getting to these answers often requires the ability to apply a specific rule-bound process, one that adjusts to each specific customer and situation.
“Emerging conversation process automation (CPA) technologies are now allowing us to build virtual agents that can have conversations that shape to each customer but stay true to rules,” Falkenberg says.
“Combining the conversational experience of Conversational AI technologies with the process rigour of business process mapping tools, CPA allows us to build, deploy and manage virtual agents that can operate at the level of human experts,” he continues. “They can have both inbound and outbound sales and service conversations across many digital channels, without the risk of getting it wrong. Every conversation can be had in a consistent, compliant and hyper-personalised way.”
Virtual agents are increasingly being used to automate the high volume, time-urgent and repetitive conversations that human agents are still being forced to have. Examples include outbound lead qualification conversations, and the conversations relating to renewals, cross sales, missed payments and collections. They are also being used to field many of the inbound service queries and disputes that clog up contact centres.
Human agents remain critical.
“Virtual agents are brilliant at handling conversations where the customer just needs something done or solved for them. However, when customers need to be heard, to have the opportunity to explain their frustration or concern, virtual agents are not the answer. This is when you need to engage with another human.
“These calls build connections and talk more to the heart than the head. They are the calls that ultimately matter,” he warns. “They are less structured, less rule-bound, and typically longer in duration – and they make a profound difference to the ultimate customer experience. And it is here where human agents really show their value.”
