What’s a ChatBot — What’s not?
IBM on the difference between ChatBots and the new Future of Virtual Assistants
Ask 10 different people what a ChatBot is and you may very well get 10 different answers. The official definition tends to be something like this: a computer program designed to simulate conversation with human users via auditory or textual methods, via the phone or over the Internet. But if you listen to pioneering cognitive experts, you quickly learn that there is a very large difference between what we know as ‘ChatBots’ — and the future of Virtual Assistants (VAs).
To a large extent, the difference experts are referring to is Cognitive. So if Cognitive Assets are what’s changing the future of VAs, let’s make sure we understand what Cognitive is. When you hear Cognitive, think ‘Pattern Recognition’. Cognitive IoT allows analytics to incorporate vast data from our ever-connected world. Cognitive Analytics then leverages machine learning to discover patterns in both structured and unstructured data. And Cognitive Computing leverages the analytics to emulate human cognitive reasoning functions.
The emergence of these Cognitive Patterns allows for, never before seen, insights and as you can image, changes the way computers interact with data. Cognitive Assets allow new depths of data exploration, discovery and decisioning. ‘I see now that when X happens, Y happens so if my goal is to lessen Y I must allieviate X. Unless there is a choice Z that comes into vision which is even more suited to lessen Y’. It is the same way a human brain would work if the person could instantaneously recall all possible information on a given problem, and look for patterns that would solve it.
Now that we have Cognitive VAs that can offer endless useful insights to support services — and sales, how do we decide on how to deliver it? Virtual Engagement may be the biggest challenge on the road to Super Assistants, and maybe the most important. After all, if your customer doesn’t like, or will not interact with your agent, virtual or otherwise, you’re dead in the water. So stakes are high to deliver the best and brightest Cognitive VAs!
Being a Virtual Agent is not easy. First, you must understand the context of the conversation — who is this person, what is the full history of interactions with the company, why might they be calling now… Next, you must use all of the information available to you to offer a personalized experience based on insight driven answers and recommendations. And then you have to learn from each experience and adapt your behaviors — a feat difficult for most humans!
That’s not all. The customer of tomorrow will want to converse 24/7 from any device they happen to pick up — and have the same conversation still going! They’re also going to want assurance — which will only come with talking to an Agent who can answer questions with confidence and seamlessly handles processes and transactions.
Well, one asks “If a Virtual Agent can do all that — what will the real agents do?” Experts answers this universal question very succinctly, ‘The goal of cognitive computing is to amplify Human Cognition — not replace it. Yes, VAs can go direct to consumers to serve, guide and advise — but there is a human gap they will never be able to fill. Would you discuss your child’s medical condition with a Virtual Agent?’
You should think of VAs not as usurping Agents and Advisors — but Empowering them. Advisor facing Cognitive Assistants gets the agent the information they need to be excellent at their job — everytime. Other Employee-facing applications are being used in the field to make workers safer, faster, better.”
And don’t think these improvements don’t make a difference — they do! Cognitive Care allows customers to begin to feel confident in their VAs, who are lessening their wait time and shortening call times to complete transactions. Currently, in a 6 minute customer service call, 75% of that time is devoted to agents doing manual research, with valued customer interaction at a dismal 25%. Initial studies in call centers are predicting a reduction in research time to 10%! The effects that could have on global organizations is mind-boggling.
The Virtual Assistants and Agents (VAs) now being built are not question and answer, corpus-based solutions. They are designed to replicate different personas a customer may interact with when contacting an organization. There are sales personas, advisor personas, service representative personas — even claims personas. Each is developed not to answer rote questions with predetermined answers, but to understand the intent of a caller — and work to meet that need, whatever it may be.
This is not a question/answer approach. VAs excel at engaging in extended, complex multi-dimensional natural language conversations. They can also understand context, such as should an answer be different depending on who is asking the question — and when.
Sentiment is an important part the pie. Emerging VAs do not make callers push 0 ten times to get a customer service representative on the line. These assistants can actually ‘sense’ and understand when the customer is feeling frustrated, and pass them off to a human, proactively. They also are aware of what types of questions should be immediately turned over to customer reps, by being aware of the topics that are not appropriate for a virtual agent to answer. An example of that may be an emergency, or an issue that demands third party involvement.
Modern VAs are also able to deal with multiple variables. When agents are asked to add a beneficiary to a policy they may come back and ask, “Is there presently a trust in your will? Was this person a beneficiary for you prior to this request? — if not we will need their approvals. And also, is yours a community property state?” The ability to ask pertinent questions, unique to the individual asking for help, is a departure from other traditional solutions you see in the market today. Many solutions max out at 3 or more variables. Our VAs deliver a deep conversational, truly omni-channel experience — and its making all the difference for our clients.
So ChatBot technology is a piece of the technology put into Virtual Agents, but it’s not the destination. VAs are provided with a wealth of knowledge pertaining to their subject area and persona, and then Cognitive Agents sort through that plethora of knowledge to serve the customer best, using natural language.
The difference between what past and present VAs is that clients can use a platform from which to build the type of virtual agent, uniquely needed by their organization at that time. Most providers in the industry today are offering specific solutions for particular needs rather than building a platform that can grow and change depending on emerging or increasing customer expectations.
These platforms and tools can do an array of things. Organizations are solving pain points in Underwriting, RFP management — even in Corporate Legal departments! Interestingly, interestingly many are seeing an incredible opportunity specifically in Insurance, and adding solutions to expand portfolios across that value chain, in particular.
Words of Wisdom
And so, as the world begins a new relationship with Virtual Agents. Here are some helpful hints to guide us in the beginning of an incredible journey.
Understanding the assets of Cognitive is a must for industry moving forward. End of Story.
And the beginning of another…
