Who is managing the digital employees?
Bots are similar to human employees in some respects; both need ongoing supervision, training and feedback. Here's what you need to consider.
24 Aug 2018
Who is managing the digital employees?
At Digital Transformation North America in September, Jerrid Hamann, Digital Customer Experience Strategist, Verizon, will be part of a panel discussing the business case for investing in artificial intelligence (AI). Read his article below for some pre-event insights.
It seems everywhere I look, someone is selling a bot, creating a bot, warning about bots or preaching the benefits of bots. Are we at the peak of the hype cycle for AI and bots in marketing and customer experience? Borrowing terms from Gartner’s hype cycle model: are we about to crest over the “peak of inflated expectations” and slide down into the “trough of disillusionment” in the coming years?
One sign of ‘peak hype’ that I have observed is that automation scripts (with no hint of machine learning) are being labeled 'bots' by vendors and executives, mostly because AI and bots are cool. Remember that ‘automation’ is scripts that follow rules programmed by a software developer. ‘AI’ software is designed to mimic human thinking, which includes seeking patterns, learning from experience, and self-selecting responses based on the context of a situation.
Serious participants in the AI world know that bots are more than automation scripts. In chatbot applications deployed in customer service environments, they adapt to a customer’s conversational remarks and deliver a modest range of answers to common questions, or properly triage a customer issue for hand-off to the right back-line agent.
In the digital customer experience sector, customer service leaders are excited about replacing scores of inbound call center agents with chatbots, but how practical is that in the near to midterm?
Customer service executives should consider these chatbot agents as ‘digital employees’. Compared to traditional employees, the chatbot army does not need a salary, medical benefits or a retirement plan. They will not take vacation, they will never get tired and they can work the day shift, the swing shift and overnight hours. They will even work weekends without complaint.
These digital employees, however, are similar to human employees in some respects: Both need ongoing supervision, training and feedback.
There is little doubt that continued investment and focus on AI and machine learning will deliver great advances in automated customer service and marketing analysis, and more broadly, autonomous transportation, expert medical systems, etc. But, over the next few years, we’ll likely see less ‘bang for the buck’, requiring continued patience for service providers and customers alike as this technology matures.
It seems everywhere I look, someone is selling a bot, creating a bot, warning about bots or preaching the benefits of bots. Are we at the peak of the hype cycle for AI and bots in marketing and customer experience? Borrowing terms from Gartner’s hype cycle model: are we about to crest over the “peak of inflated expectations” and slide down into the “trough of disillusionment” in the coming years?
Is it automation or AI?
One sign of ‘peak hype’ that I have observed is that automation scripts (with no hint of machine learning) are being labeled 'bots' by vendors and executives, mostly because AI and bots are cool. Remember that ‘automation’ is scripts that follow rules programmed by a software developer. ‘AI’ software is designed to mimic human thinking, which includes seeking patterns, learning from experience, and self-selecting responses based on the context of a situation.
Serious participants in the AI world know that bots are more than automation scripts. In chatbot applications deployed in customer service environments, they adapt to a customer’s conversational remarks and deliver a modest range of answers to common questions, or properly triage a customer issue for hand-off to the right back-line agent.
The chatbot army
In the digital customer experience sector, customer service leaders are excited about replacing scores of inbound call center agents with chatbots, but how practical is that in the near to midterm?
Customer service executives should consider these chatbot agents as ‘digital employees’. Compared to traditional employees, the chatbot army does not need a salary, medical benefits or a retirement plan. They will not take vacation, they will never get tired and they can work the day shift, the swing shift and overnight hours. They will even work weekends without complaint.
Digital employees (sort of) have feelings, too
These digital employees, however, are similar to human employees in some respects: Both need ongoing supervision, training and feedback.
- Frontline supervisors need to ‘listen in’ periodically on chatbot conversations with customers to ensure the bot’s language and attitude reflects the brand’s image and desired experience.
- Back-line expert agents must regularly evaluate the chatbots’ answer accuracy and be ready to handle escalated issues/questions when the chatbot gets off-track or out-of-depth.
- If the chatbot is handling sensitive customer information (particularly in a regulated industry like finance, healthcare, or telecommunications), legal and regulatory specialists must periodically review the bots’ system access and data handling. Your company probably has existing rules and procedures for granting (and rescinding) your human employees' access to sensitive systems and data. Consider registering your digital employees in the same way using the same procedures.
- Training and development staff must train your customer service staff with new information upon the release of new features or entirely new products. For human employees, this might be a one-hour or one-day training class. For digital employees, this training will be some form of offline machine learning (uploading new data sets to change the AI model) or online machine learning (continuously feeding the model new information).
- Finally, the customer experience team must review CSAT (customer satisfaction) and other feedback loops from customers after interacting with a chatbot to ensure the overall chatbot army is delivering an experience on par with (or better than) the traditional human-staff contact center. If CSAT scores are slipping, customer experience professionals must dig into the details to determine where to improve the chatbot experience.
There is little doubt that continued investment and focus on AI and machine learning will deliver great advances in automated customer service and marketing analysis, and more broadly, autonomous transportation, expert medical systems, etc. But, over the next few years, we’ll likely see less ‘bang for the buck’, requiring continued patience for service providers and customers alike as this technology matures.