Artificial Intelligence (AI) has been a favourite topic of science fiction adventures for several decades. In the make believe world of Hollywood, AI is usually linked to a comment on our humanity or machines turning against humanity, such as The Terminator or 2001: A Space Odyssey.
In the real world, the field of AI research has been around since the 1950s. In these early days, very basic functions such as computers playing and winning at checkers was deemed AI. Since then, the definition of AI has evolved along with the ongoing expansion of computer functionality.
Today, AI is primarily applied to machines that mimic the cognitive abilities of the human brain, such as problem solving and learning. Microsoft’s chatbot Tay is a very recent example of a machine that can learn. Microsoft unleashed Tay on the world, the latest and greatest mix of AI and machine learning. Users globally were able to interact with Tay via Twitter and messaging apps like Kik and Group Me. Unfortunately, within 24 hours Microsoft’s project was hijacked by a number of users who used disturbing content during their interactions with Tay. Although Microsoft had no choice but to shut down Tay, this short lived experiment was still deemed a success in that a machine was able to learn, albeit a very unpleasant learning.
The integration of AI into our daily lives will continue as this technology evolves through its maturity curve, with Apple’s Siri as one of the most well known AI applications used by many on a daily basis. Although there are many more examples of AI, the world is about to experience an explosion of AI applications courtesy of Facebook.
In April 2016, Facebook announced that third parties can use their Messenger application to create their own chatbots (short form for chatter robots – a chatbot is a computer program that uses AI to simulate conversation with humans), enabling organisations to automate online conversations with their customers. Facebook Messenger’s 1 Billion plus users will bring AI and chatbots into the mainstream in a big hurry. So if you haven’t already had a conversation with a computer, you soon will.
So, how will this new frontier affect the future of contact centres? Will chatbots fully replace the need for human powered contact centres?
Short answer is no, not yet at least.
Chatbots will not only reduce the volume of transactions for contact centres, but they will also eliminate most of the simpler transaction types. Similar to the effects from other technologies such as Interactive Voice Response (IVR) and smartphone/tablet apps, contact centres will experience a surge in higher complexity contacts as a percentage of their overall contacts. As chatbot technology continues to learn and evolve, the most complicated and difficult transaction types will be left for human powered contact centres to manage.
Here are the top five implications that contact centre leaders must consider as chatbots take hold over the next few years and beyond:
1. Some metrics may behave in undesirable ways
Contact centres are typically awash in a sea of metrics and stats. In addition to understanding what has happened, various metrics are used to forecast and predict future demand, performance and capacity. As contact centre leaders prepare their long-term strategies, they should consider some of the following:
Average Handle Time (AHT)
This is a fairly obvious one in that as the volume of simpler contact types (which typically carry a lower AHT) reduce over time, the overall AHT for the centre will increase.
First Contact Resolution (FCR)
An ongoing increase of difficult and challenging contact types may lead to an increase in the number of activities an agent must perform to resolve an inquiry, such as the need to gather more information and/or liaise with colleagues. The net result is potentially a steadily declining FCR over time. A way to mitigate this risk is to empower the agent with advanced tools and support (see point #3).
Net Promoter Score (NPS)
NPS scores are typically only sent to customers that have engaged with a contact centre. The danger of this approach is that the results are only representative of a proportion of your customers. As chatbots remove most of the simpler contact queries from contact centres, NPS results from these customers will most likely only represent a sample of your customer base that has a highly complex inquiry or an escalated complaint. While this information is valuable, it is also worth considering how to capture survey results from customers that interact with your other channels, such as your knowledge hub or virtual agent tool to get a true view of your NPS.
2. Contact centre agents will be the ideal teachers for chatbots
Similar to the current experience with IVRs, customer-chatbot engagements that are unable to progress will be sent to an agent to review.
The agent will determine whether they: