In a relatively short timeframe, automation has risen to the top of the agenda of customer service organizations. Robotic process automation has become one of the five most searched topics on gartner.com. The firm also predicted that one in four customer service operations will use virtual assistants by 2020.
The attractiveness of automation is not new. It drove the adoption of Computer Telephony Integration (CTI) and Interactive Voice Response (IVR) in the past. It is also a logical consideration for organizations pulled between rising volumes of interactions, the mandate to provide better experiences, and the need to improve their efficiency. The cloud and Artificial Intelligence (AI) have paved the way for new options. Let's explore them.
Virtual customer assistants
Virtual Customer Assistants (VCA) have enjoyed a lot of coverage. Natural Language Processing (NLP) has enabled a new form of interactive—conversational—dialogue. Machines can be trained to understand intents expressed by customers using their own words and prompt for the missing information. VCAs promise to address the shortcomings of rigid voice self-service applications and site searches returning too many results.
The category has become crowded. I am now tracking over 75 providers. In 2017, Forrester proposed a vendor selection for enterprises but the market is still in its early innings. While most solutions are built for digital channels, a growing number of companies such as [24]7, CallDesk, Interactions, Nuance, Omilia, and SmartAction have been going after voice applications.
The technology does work and results can be impressive. However, expectations tend to be overset. It led Gartner to predict that, by 2020, 40% of applications launched in 2018 will be abandoned. Forrester has also been warning the industry about the devastating impact on the customer experience of poorly performing chatbots. Enterprises are finding the implementation harder than they thought. At this stage, the most successful solutions are the ones focused on specific use cases and industries.
Robotic process automation
Robotic Process Automation (RPA) is a much more mature technology. RPA uses software robots to mimic the interactions of a human with an application. It has been maturing for more than a decade and is now inside a tornado. Early leaders include Blue Prism which became public in 2016, Pega further to its acquisition of OpenSpan, Automation Anywhere, and UiPath which raised each a half billion dollars.
The technology can operate autonomously (unattended) or side by side with a person (attended). The latter makes it very relevant for customer service. RPA also provides a simple integration method with applications: it doesn’t need any development and can accommodate legacy applications that don’t offer APIs. Eventually, RPA can be layered on top of existing software stacks, put to work with little IT effort, and deliver rapid results using an agile approach.
These attributes should make RPA very compelling for customer service. However, unattended use cases in the back office for finance and shared services have enjoyed the most traction. They have bent the industry trajectory. Many vendors are ramping up their platform capabilities to support large deployments. Indeed, large pools of software robots require supervision and oversight. Vendors also shifted their focus to getting IT organizations to set up centers of excellence. This has left the customer service market poorly aware of the possibilities of RPA and served by a subset of specialists such as Jacada, Kryon, or NICE.
No-and-low-code platforms
No-and-low-code platforms are the latest iteration of Business Process Management (BPM) software. BPM platforms, like RPA, have been available for many years. They let you rebuild and automate a process by orchestrating all its steps across applications and humans. BPM projects require pretty intensive IT involvement and effort. In particular, the integration with applications is often tedious and takes time to validate.
The next-generation platforms either called no-code or low-code promise shorter development cycles and faster return on investment. It is making the technology a compelling option for customer service automation. Two key players, Pega and Appian have already created offers for the contact center market.
Workflow technologies are not new in the space. Contact center providers such as Avaya or Genesys and CRM vendors such as Salesforce with its Lightning platform have added capabilities to their platforms. We need to mention ServiceNow, which built its platform around the management and automation of workflows and broke into the customer service market a couple of years ago. The technology is getting simpler and new entrants such as Bright Pattern or Talkdesk have started to offer solutions that customer service organizations can leverage by themselves.
Bot platforms
The latest option for automation is bots. Bots are specialized and can exist in many forms. For example, an answer bot can provide answers to frequently asked questions. It is similar to a virtual assistant but instead of being fed by chat or voice transcripts, it uses a knowledge base and can be trained through unsupervised learning. Representative providers include Alterra, Helpshift, Microsoft, and Zendesk which coined the term.
Bots can be used for a multitude of tasks. Several vendors have assembled bot development platforms for customer service. Some such as Ada or Passage AI are pure bot development platforms while others from the likes of Helpshift or Intercom are integrated into their customer service solutions.
With so many options blossoming and customer support organizations pressured to do more with less, automation seems inevitable in the space. It has triggered angst with some pundits predicting it could impact up to 30% of the customer service activities. Together with Sheila McGee-Smith, we will discuss all these topics at the "Automation vs. AI-Assisted Humans: Where to Draw the Line" panel at Enterprise Connect. Join us there to continue the conversation.
This article was originally published on BCStrategies