Putting human call center agents on hold

Published by reposted only Date posted on February 23, 2017

By Joseph Williams, Feb 23, 2017

How long do you think it will be until virtually all tier 1 call center agents are replaced by AI bots? I addressed this a bit in my Machine Learning session at last year’s Business Communications Summit.

I’ve been discussing the topic with my colleague Dave Michels, and we agree that the impact of artificial intelligence on contact center operations will be significant. Chatbots are the first step, but automated bots that naturally interact with callers will initially reduce and eventually eliminate all but the most complex human agent interactions during the next decade.

The industry has arrived at a formidable convergence of natural language processing, speech synthesis, machine learning, and declining compute costs that will make it possible to produce outstanding customer service outcomes from AI bots within the next five years – if not sooner.

There is a powerful motivation here; there are some 2.2 million call center employees in the US and another million or so supporting US customers offshore. These call center employees occupy more than five million square feet of office space. Pencil out the costs to industry to provide customer support (don’t forget to add contact center software, office furniture, and headsets) and you arrive at tantalizingly large expenses that will be targeted for reduction by vendors of new robot customer service representative (CSR bot) technologies.

CSR bots are already here in rather primitive forms for call automation as in auto attendants and virtual receptionist services. These are largely call routing and simple information (e.g., account balance or bill due date) services. The coming wave will enable a CSR bot to replace a significant number of human CSRs and answer, in completely natural language, a considerable array of customer calls.

The CSR bots will likely provide a better overall customer experience in the aggregate than do human CSRs [a chilling idea, I concede, but no less true for being chilling] and will do so consistently 7x24x365, probably in a cloud environment where enhanced interaction options will be available via microservices. These CSR bot applications will instantly scale to handle any call volume in any language across pretty much any device (voice, video, text, or perhaps even in virtual reality) at a fraction of today’s costs. These CSR bot solutions will enable firms of any size to offer sophisticated customer call center services.

The trend towards automation is already accelerating among mobile collaboration vendors. Recently published research predicts that knowledge-delivering chatbots will be standard offerings in mobile collaboration suites by the end of this year. Chatbots are computer programs designed to produce conversations with humans, typically through some kind of text-based interface. The voice extension of this technology has been to Amazon’s Echo (aka, Alexa) or to Microsoft’s Cortana. While early chatbot prototypes were clunky they are rapidly improving helped by an explosion of powerful tools (IBM’s Watson, Octane AI, gupshup, and many others).

The AI breakthroughs that are enabling chatbots are already being applied to the call center. Virgin Trains in the UK is using the Celaton system to handle many customer inquiries while cleverly routing more complex problems to human operators, all the time learning from each call. One of Europe’s best-known analysts, Sarah Burnett of the Everest Group, has already predicted that CSR bots are going to get so good that CSR jobs are going to vanish for repetitive tasks. It is coming, folks.

In 2017, Dave and I are going to work together on providing analysis of emerging CSR bot products and technologies. We are going to look at what traditional UC vendors are doing and what potentially disruptive market entrants might come into play. If you have ideas and comments you want to share on this coming disruption, please do so below or by dropping me an email.

Get Email from TUCP
Categories for Archives articles