With the advent of faster communication platforms across devices, today’s customer is willing to connect with service providers through new channels (social media, self-help, online assistance, live chat etc.) and seek a more seamless and simpler customer experience. Customer also demands a faster grievance capture. Here we look at the latest inventions in the area of machine learning and natural language processing and the benefits a communication service provider can derive from these while attaining efficient handling of customer's requests. Thereby improving customer satisfaction.
Typically, a customers’ need to interact with a CSR (customer support representative) of a telecom service provider can be categorized around areas like - Request for New Service / Change of Service/ Service Termination, Complaint resolution, Status update on tickets raised or Bills and payments related queries. There are organization defined business processes for handling each type of customer requests. CSRs follow these business processes while interacting and collecting details from a customer.
With the advent of IVR (Interactive Voice Response) some of the activities were moved to machine interaction thereby reducing call volumes to the CSRs. However, due to technical limitation of IVR, customers can interact with the system only via key presses and not in their natural language. To keep the key presses minimal, only few services are provided on IVRs. Such as Bills payable date, Last billed amount, Status of last complaint, Start/ Stop certain value added services (VAS).
Latest innovations in natural language processing and artificial intelligence have provided machines the ability to mimic human conversations which can further enhance customer service experience and allow service providers to reduce their call volume by a great extent thereby reducing their operating expense.
Welcome the BOTs
Personification of machines became popular with the advent of personal digital assistants, starting with Siri and later Google Now and Microsoft Cortana. While these digital assistants are capable of doing a lot, there is another category of digital interfaces – the Chatbots, which can hold human like textual conversation in order to accomplish a task.
“Imagine texting a number to order pizza and having it delivered without ever talking to a real human -- that's what bots are all about. Specifically, a bot is an application that performs an automated task, such as setting an alarm, telling you the weather or searching online.” - Sarah Mitroff, CNET.com
The overwhelming growth of messaging applications like SnapChat, WhatsApp and Facebook messenger have made the world adopt texting as a natural means of communication. Simplicity of chat interface coupled with high speed and the plethora of add-on features (for e.g. ability to attach images, sound, video, emoticons etc) has made messaging apps the preferred choice of correspondence. Chatbots intend to capitalize on these messaging interfaces and be added in your contact list so that you can start an interaction in a jiffy, anytime of the day. In short, they aim to be the acquaintance who knows all about a subject and is eager to help.
Chatbot for Telecom Customers
Some telecom service providers have started to provide chat based assistance on their websites. Currently these are handled by human support agents and are also limited in functionality.
Pattern of correspondence between customers and support agents
• Customers begin a chat to seek information or lodge a complaint.
• Customer inquiry is addressed by support agents by following a specific workflow.
• Customer requests and complaints are repetitive in nature
Hence there exists a huge potential to automate customer's request handling and an opportunity for the telcos to effectuate chatbots and reduce expenditure.
Example chat session with chatbot
1. Personalized greeting to customer based on gender, time zone and preferred language. Details from backend system are retrieved based on mobile number.
2. Seek information about the problem. Let’s say customer need plan details.
3. Bot can either provide a link to plan detail webpage or query CRM system to share any customized plan details as a document. Bot can also offer to lodge a service request. Since the systems are pre-integrated ticket logging is faster.
4. Share ticket number with the customer. Customer also gets an SMS. Wish and closes the session.
This results in a happy customer due to effective and efficient request handling. At the same time, telco saved effort of its support agent who can focus on tasks requiring manual intervention.
Possible areas for Chatbots to handle
• Billing and Payment - Queries on bill plan, bills payable, due date, payments received, outstanding data/voice limit, applying discounts, changing bill plan and other routine queries.
• Complaint Handling - Status updates, lodge complaint, complaint history, schedule technician visit etc.
• Faults – L1 support, create trouble tickets, inform about network outage etc.
At any point, where a the bot cannot handle the interaction it can either create a trouble ticket and ensure a callback from customer support agent or if the platform permits, directly initiate a voice call and transfer to customer support.
Communication service providers around the world are investing in ways to evolve and extend customer service channels by hearing what their customers have to say via all possible means of communication. The below figure depicts a paradigm shift in delivery of customer service.
In such a multi-channel scenario, chatbots can really form the first line of customer support and can provide a true Omni-channel customer experience across multiple chat based platforms and devices e.g. SMS, Online Messing applications, Chat platforms on web pages, Messaging apps on smartphones.
The Ground Work
Define Primary Use Case - "every chatbot will provide a specific service. Stating the purpose of its existence is important to identify its behavior."
Knowledge Base – Start with listing down all possible service specific questions and corresponding answers
Structure data - structure the information into relevant categories and subcategories so that the bot can answer top-level questions and then dig deeper into a subject.
Build Intelligence – Bot needs to be continuously trained to understand what the user types, pick relevant phrases, match it with the answers in its knowledge base and respond.
Benefits at a Glance
Increase Revenue - By communicating personalized offers and Assisting in customer behavior analysis
Reduce Operating cost - Providing low cost, scalable and robust solution for customer support.
Increase Retention - Omni-channel experience and Personalized customer care.
A Step Further
Chatbots can initiate a dialogue with customer on defined triggers like – Approaching bill due date, Usage limit met or Propose to switch to available wifi hotspot in vicinity when available. Chatbots can also utilize data analytics about consumer behavior and usage pattern and assist in upselling. For e.g. Propose higher limit plans on detecting high data usage or Propose roaming package on detecting user outside home network or propose Special top-ups and offers.