That Person Responding is Not a Person Anymore!

By Bennet James Bayer, Technologist, Board Advisor, Former Global CMO & VP Strategy with Huawei, Stealt

Bennet James Bayer, Technologist, Board Advisor, Former Global CMO & VP Strategy with Huawei, Stealt

A few years ago Gartner predicted that by 2017 25 percent of online identities would be non-human and further that this number will triple by 2019. Chatbots, digital agents and Avatars are becoming a part of the enterprise digital transformation impacting development of both internal enterprise user interactions and external customer relationships.

Having spent most of my career building emerging technology solutions, I started working with natural language in 2000 and for some, the core technology goes back to the first artificial intelligence (AI) attempts in the late 60’s and 70’s. I implemented large-scale chatbots, agents and even Avatars in a number of customer solutions and with my employer at the time.

"The uses are endless; combined with search technologies chatbots blend with mission-critical applications and customer interactions"

Many chatbot producers claim AI to automate the interactions; however, they are mostly Boolean word phrases in pattern matching of a database with programmers manually completing the phrases, but it is improving rapidly.

A chatbot is an application emulating a human-like response using a speech or text response. If this were a contact center I would be a digital agent and I would be asking, “Would you like to know more?” Chatbots can serve as a form of personal assistant, butler and concierge or help agent as we see in Siri, Alexa, Google Home and similar market solutions; most Chatbots though are rather simple.

When the bot is autonomous, responding to an inquiry or perhaps responding to an online or network inquiry, they are often much more complex and are referred to as an Avatar.

The uses are endless; combined with search technologies chatbots blend with mission-critical applications and customer interactions. Do you have a Skype account? There are now hundreds of simple chatbots available ranging from Expedia travel information to answers to your medical questions.

Chatbots can be either reactive or proactive. Reactive bots are often selected by the user (as in the Skype example above) from a menu, mobile app or web site. Another example is “social listening” found on Facebook, Twitter and others site, even in response to SMS.

Proactive bots are more complex and add intelligence and should respond in real-time with prediction of the user/customer intentions. These provide specific assistance; think of shopping for running shoes and the bot offers suggestions as to options.

For me the complexity of a box is around its use and complexity; they can be:

• Institutional–using natural langue to respond to questions raised.
• Personalized–provides user-specific responses by connecting to enterprise applications, databases and likely offers a menu of options which are remembered by the system.
• Transaction-based–helps the user complete a series of steps, likely with integrated customer data (CRM being the most common).

Rational for deployment ranges from reduction of operating costs, improving customer expectations and experience, increasing the volume (because they scale so easily deployment in contact center solutions is quickly justified) and market differentiation.

Similar to implementation of Unified Communications, I suggest deploying bots on a use case by business unit or department. Some of the potential benefits include:

• Customer Care/Support. Providing scale handling routine questions, support and tasks which frees care teams for more complex issues.

• Sales and Marketing. Lead generation, online response and blending with content marketing, revenue generated with existing clients. This is also where Avatars will be implemented for mobile or online search responses. An Internet subject inquiry triggers the Avatar to respond, assisting the customer and ideally directing to the companies web page.

• Channel Programs. Assisting partners find relevant information to improve their customer performance and response.

• Customer Engagement. Conversations provide far more specific demographics and analytics for marketing technology systems with scale. This creates cross-organization intelligence and tracking providing a shared view of each customer and feeding martech RBDMS.

Internal use improves collaboration and in large global organizations the ability to “find” information is greatly enhanced. Ultimately this requires integration with all mission-critical systems. However, security is a critical component and a rules-based hierarchy with tracking is recommended. You might not want every employee to ask for salary info or what product is the most profitable by region and country. But the technology is already finding success in communications, utilities, retail, healthcare, travel and finance.

The more complex the chatbot the more likely you are to expect data output from the interaction. Data helps predictions of customer interactions and it is, like most technology, a learning journey. Machine learning is improving and leading toward true AI but the improvements in advance analytics and insights are quickly realized.

To consider deploying bots I look at construction of voice mail systems. First one asks when is a human required and when should a chatbot escalate a conversation to a live agent.

There are a number of chatbot solutions available which can either be customized or provide a basic platform. Each IT group needs to review their time, skill set. Some of the disadvantages of building your own include lack of analytics output, time to deployment, higher cost, customer development (looking at where you want this tech to be in five-year’s time) and risk of failure.

Start by defining a business case. Will it reduce costs, support growth without increasing cost, monetized self-service and/or improve customer loyalty?

A simple checklist for a chatbot solution should include:

• Does it solve specific customer/user issues?
• Can it contribute data to the entire customer life-cycle?
• Can it learn and predict user intentions over millions of transactions?• Can it personalize the conversation and improve experiences as a part of the customer journey (note, you need to map the customer journey to address this).
• Does it have the ability to escalate?

Chatbots are a growing part of the digital business future. It is a journey. It is a conversation, one where you never exactly know “who” you are speaking with.