Remember when we weren't sure if self-service knowledge bases would work? Or if they would only work for certain populations? Today, my mother's best advice on many topics is to "Google it." My, how things have changed.
In the late 1990s and early 2000s we saw significant returns on investment from customer service organizations deploying self-service knowledge bases to deflect calls away from their contact centers (mostly calls, at that point), reducing contact volumes initially by 30 percent to 40 percent. In the past few years, contact centers moving toward omnichannel have seen increases in efficiency by bringing social into the mix, driving further deflections from higher-cost channels and enabling multichannel agents to manage more cases at the same time.
However, chabots, when effectively deployed, promise to cut more costs out of contact centers through greater deflection from agent contact, freeing up agents for higher-level escalations, and ultimately accelerating case resolution while driving greater customer satisfaction. The quality and consumer acceptance of chatbots is on the rise, for a number of reasons.
First, innovations in chatbots, where a bottom-up approach (analyzing large volumes of contact center records to understand customer support requirements and bubbling up the appropriate dialogue trees) instead of the top-down model currently in use (where admins or business users script what they expect customer case conversations to look like) will deliver a more broadly usable automated vocabulary and better results.
Second, as more consumers adopt text, SMS, and other channels as their support channels of choice, bots are better able to manage those cases and multiple cases at the same time, while the data culled from those interactions can be rapidly integrated into training of future bots. When voice is not the primary channel, agents are decreasingly the most efficient option for resolving multiple cases.
Third, bots don't call in sick. As employers grapple with the changing economic, immigration, and tax landscape, the appetite for hiring and training in high-turnover areas, such as customer service and support, is already low.
Don't get me wrong, for conversational interactions human beings will be the contact agent of choice for some time to come. Agents are not going to be replaced wholesale by bots overnight, but planning a bot strategy now makes sense. Many innovators are investing in applying artificial intelligence and machine learning technologies to bots, and bots are likely to a great differentiators and budget savers for those in service who get it right.
So what should we be thinking about today? First, look at bot technologies that follow the bottom-up approach to leverage the intelligence of AI and text analytics to build more contextual conversations based on real case interaction histories. Second, rethink agent scheduling and training so they can be complemented by bots and refocused on higher-level interactions as bots clear more lower-level cases. Third, look at new business models for bots beyond just case completion and lower-level inquiries that could drive revenue beyond just lowering service costs.
Rebecca Wettemann is vice president at Nucleus Research.