Chat Bots are taking over the responsibility of handling customer support, talking to leads and in many cases even converting them into sales for your enterprise. These bots are now the de-facto standard for any enterprise looking to maximize and harness online handles to improve productivity and conversions.
While they are all the rage, writing them and actually making them work needs a lot of effort and subsequent maintenance to get them setup and keep going. They also need to have very good NLP and artificial intelligence capabilities built into them to function optimally. What was once a herculean task has now been simplified by IBM Watson. The IBM Watson Conversation component makes writing and commissioning these bots very easy and achievable in a short time.
Now you can add a natural language interface to your application to automate interactions with your end users. Common applications of this include virtual agents and chat bots that can integrate and communicate on any channel or device.
Watson Conversation combines a number of cognitive techniques to help you build and train a bot – defining intents and entities and crafting dialog to simulate conversation. The system can then be further refined with supplementary technologies to make the system more human-like or to give it a higher chance of returning the right answer. Watson Conversation allows you to deploy a range of bots via many channels, from simple, narrowly focused Bots to much more sophisticated, full-blown virtual agents across mobile devices, messaging platforms like Slack, or even through a physical robot.
With this it becomes very easy to
- Add a chatbot to your website that automatically responds to customers’ most frequently asked questions.
- Build Twitter, Slack, Facebook Messenger, and other messaging platform chatbots that interact instantly with channel users.
- Allow customers to control your mobile app using natural language virtual agents.
You need to give structured inputs about domain expertise in the form of intents, entities and crafted conversation. And then the service gives you a trained model that enables natural conversations with end users
Many large corporations have used this to create effective chat bots. Let us look at one example of Autodesk. Bogged down by routine yet complex issues with no way to prioritize urgent cases, Autodesk needed to handle increasing volumes and boost customer satisfaction without expanding its workforce.
As Autodesk makes the switch from a licensing business model to a subscription-based model, it is reaching a larger, more diverse customer base—one that increasingly expects frictionless self-service options. But transforming customer service is not a simple matter. Though most inbound customer interactions are related to routine issues such as activation codes or account changes, they still require an agent with extensive knowledge of Autodesk’s complex systems and processes to chat with the customer and do some detective work. It’s not work that can easily be automated or captured in a list of frequently asked questions (FAQs). With rising volumes and no way to triage the long queue of support cases, the contact center was facing average resolution times of 1.5 days—much too long for customers who depend on the software to do their jobs. Autodesk needed a way to handle the increasing traffic while reducing costs and meeting customer expectations for faster, more convenient support.
Autodesk is using cognitive computing enabled by IBM Watson conversation to divert inbound traffic away from the contact center and give customers what they want: a simple and free flowing self-care solution supported by a dynamic and intelligent chat bot. Using natural language processing, the solution converses with customers to understand the source of their problem, parsing not only the literal meaning of customer input, but also interpreting the intent behind it. The system was trained to recognize and categorize common customer issues and solve them on the spot if possible—for example, it can issue a new activation code by collecting the right information from the customer, pinging the associated systems and providing a new code within seconds. For more complex and uncommon issues, the bot can create a support case, prioritize it and route it to the right agent. Continuing to learn from repeated interactions with customers, the cognitive solution becomes increasingly accurate in diagnosing problems. The solution is currently in its pilot phase, but Autodesk plans to expand its ability to resolve issues directly—without human intervention—by incorporating learning content such as training manuals, technical support documents and videos into the corpus.
With this implementation, they were able to realize the following changes
- 90% lower support costs by diverting automatable cases away from the contact center
- 99% lower resolution times from hours to minutes99% lower resolution times
- Drastically increased customer service capacity to handle growing volume without growing the workforce
So all in all IBM Watson Converation is a productivity and support efficiency enabler. I am very sure that going forward, many large scale enterprises will adopt cognitive technology to improve interactions and productivity. IBM Watson conversation is correctly positioned to address this gap and take the AI mantle forward. I will be watching this space keenly for developments in this direction.
As a first step to build your chat bot using the Watson Conversation Service, sign up for FREE today on the IBM Bluemix platform. Bluemix is the cloud platform for open source and is currently the platform that you would have to use to create your chat bots https://console.ng.bluemix.net/registration/
