As consumers, our only encounters with AI in customer service are likely to be through chatbots. However, the potential for AI to impact customer service goes much deeper than just automated chat responses. For customer service agents working in a contact centre, AI has a whole host of uses. It can be used to avoid or deflect simple questions, intelligently route tickets to correct team members and speed up resolutions by offering potential solutions.

Despite the real ways in which AI can improve the customer service experience, there are still many criticisms hurled in its direction. These criticisms are partly due to a lack of understanding around the nature and role of AI. It’s therefore important to understand what AI can and cannot do in our world today.

What AI cannot do
In its early stages, AI, much like humans, needs to learn. It does this by consuming vast quantities of data. The greater the volume and quality of data, the better the machine is able to learn. The first step, for companies looking to develop AI solutions, is to ensure they are collecting and adequately tagging their data. AI cannot learn without a human teacher and large pool of data.

While AI lacks human empathy and emotional intelligence, it has the right complementary skills for human assistance. Returning to the chatbot example, from our own research at Freshworks, we can see that B2B chatbots can currently only respond to 40 percent of questions and of this percentage only 8-25 percent are successfully handled. This is a good start but it’s premature for contact centres to route all their online chat queries through a bot. We need to instead think about how AI can best supplement existing staff.

“AI brings speed and cost efficiency whereas humans provide empathy and emotional intelligence.”

What AI can do
AI is very good at automating repetitive tasks, and as a result deflecting simple questions is something AI does well. For example, a consumer questioning where their Uber is doesn’t necessitate a human response; Using the data available, AI can easily answer this. We have also built a similar tool, providing one customer with a chatbot for ticket deflection – it can answer simple questions, such as: What’s my wallet balance?”

AI can also offer a level of speed that is not humanly possible. This has a variety of benefits. Automated systems can provide customer service agents with recommended courses of action, based on previous encounters that the AI has observed. This allows the agent to offer solutions which effectively preempt additional issues that the customer is likely to experience. For example, if the offer of an upgrade or a related service has diffused a potential customer clash during a similar situation in the past, AI will know this and suggest the appropriate response. Agents can draw on this knowledge and use it to their advantage.

Speed also equals cost. The quicker a service agent can resolve a customer issue, the quicker they can move on to the next one. This leads to a more efficient contact centre and saves money in overheads. The added flipside of this is that a speedy response to a customer query is likely to translate into a positive brand experience.

Allocating resource is another area where AI excels. Often in a contact centre, tickets are raised and forwarded to the wrong team, or to an agent at full capacity, resulting in lost tickets or delayed resolution. In today’s world, many customer interactions begin on social media, however, the sheer amount of social data can make it tough to target the tweets worth responding to. Here again, AI can analyze the components of certain tweets and route them where necessary. By intelligently routing tickets, based on machine learned principles, AI can maximise contact centre capacity and effectively manage a team.

In short, AI brings speed and cost efficiency whereas humans provide empathy and emotional intelligence. The customer service industry, once it understands the limitations of AI, can use the technology to provide greater value, however, at this stage, AI should not be used to replace human-to-human interaction. Instead, AI can only be used as a filtering system, to address simple repetitive tasks, allowing humans to tackle the more complex issues, all the while drawing on the computing power of AI when necessary. This won’t always be the case – as we teach the AI engine more, its capacity to help us will increase and the tasks that it cannot do will surely dwindle.

ABOUT SREELESH PILLAI//

Sreelesh Pillai is the General Manager, Australia of Freshworks.