How Can You Build a Smartphone Support Chatbot?A handy Smartphone Support Chatbot development guide for mobile operators and MVNOs
Customer service agents at mobile operators’ and MVNOs’ call centers spend dozens of hours in solving customers’ mobile device setup problems, and instructing them on using their handsets daily.
This is no surprise for those who know how complex mobile devices and wireless networks are. Mobile operators’ call centers are unfortunately where one finds a tangled technological web. Service agents have the impossible task of fixing all service or device problems thrown their way, all done remotely via a mere phone call!
There can be multiple reasons for the increasing need for call center support in smartphone usability, setup and configuration issues: these include that modern smartphones are difficult to use, device setup often requires the configuration of several parameters on device menus, Chinese-made low cost handsets and counterfeits are common, and in some markets such as Africa and Latin-America, there are many smartphone first-timer users.
Chatbots are an ideal Smartphone Support ChannelThanks to the many chatbot platforms available, any company can now build a chatbot in no time! However, automating a customer service bot that scales to meet the needs of millions of customers can be challenging, without an experienced partner and AI powering your bot.
Smartphone owners can encounter dozens of different kinds of problems, and the solution to their particular problem often differs, depending on the device and operating system involved. Nowadays, call centers need support content for hundreds, if not thousands of solutions, depending on the particular problem in question.
Chatbots are ideal in these environments, due to their ability to provide support to identify the customer’s specific problem and to provide a personalized solution directed at that particular problem.
Chatbots are becoming the preferred platform to deal with customer care interaction in many industries. It’s no surprise why – simply by using a simple user interface, bots can easily deliver answers and solutions to problems.
Building a Smartphone Support Chatbot – Key Considerations!This article provides mobile operators and MVNOs with the key considerations and tips on how an existing Online Smartphone Self-Care Service on a website can be enhanced with a bot interface, or how to build the bot from scratch.
What is a self-care Chatbot?A chatbot is a software program for simulating a conversation with a customer service or call center agent using rules or artificial intelligence. Users interact with the chatbot via a conversational interface through written or spoken text. Chatbots can work through messaging platforms like Facebook Messenger and Telegram, or simply on an operator’s or MVNO’s website, and help to provide end-users help for various purposes – such as topping up prepaid balance, setting up an Access Point Name (APN) or email account, flushing a voice mailbox… among many other problem or service situations.
These are the Benefits of Self-Care Chatbots in TelecommunicationsChatbot based customer service benefits both operators and MVNOs in several ways. They can increase revenue by assisting in analyzing the customer behavior and can also communicate customized offers, which in turn will lead to add-on sales transactions.
Bots reduce operating costs – They are a perfect way to provide a low cost and robust customer support 24/7.
Through bots operators can increase customer retention by providing a quick, personalized customer service and an omni channel experience. Bots are ideal for dealing with repeated issues, thereby allowing human attendants to concentrate on more demanding cases.
Development strategy – A Bot Framework or a Bot Platform?Choosing the right development strategy is one of the first decisions to make. Operators and MVNOs can build customer service bots in two ways.
Using a Bot Framework, requires software development resources, but they make the development faster and remove much of the manual work involved in building bots. Bot Frameworks are supposed to abstract away much of the manual work involved in building chatbots for several communication and messaging platforms and SDKs.
There are numerous tech companies and startups who provide generic Bot Framework offerings: these include Facebook’s bot engine Wit.ai, Microsoft bot framework, API.ai, Kik and Chatscript.
On the other hand, Bot Platforms, such as Ada.Support are complete online ecosystems, on top of which chatbots can be deployed and operated. Many include a pre-integrated partner network. Bot Platforms do not require software development skills. There are several options for platforms: Motion.ai, ChatFuel and Octane AI among many others.
Ada.support differentiates in the Bot Platform space becasue of its strong track record in working with leading telecom operators.
How much Intelligence is Needed?The chatbot experience is only as good as the machine learning that powers it. It’s important to find an AI-powered platform that provides not just instant, but also intelligent support.
For example, when a user asks a bot for help in repairing the Internet connection, the chatbot must understand instantly what the user wants and provide the right response.
In this case, the bot must first ask the user to state his or her handset model (unless the implementation extracts that information from the HTTP header or IMEI code), then provide instructions for the most common root-cause for a broken mobile Internet connection, namely how to configure the Access Point Name (APN) on that handset model.
Natural Language Processing (NLP) makes this possible. As with so many of these technologies, an operator or MVNOs does not have to implement NLP algorithms on their own. There are NLP APIs available that can be used to infuse NLP capabilities into a self-care chatbot. Some of the sources for a NLP API include Wit.ai and Api.ai. Ada, a global leader in AI-powered customer service allows non-technical teams across telecommunication companies, to build and deploy a scalable, intelligent customer service bot in less than a month, and this is available in more than 100 languages.
Ada has enabled businesses to automate up to 70% of their customer service, with more than 80% recognition by using its AI powered platform, Ada.support.
Ada’s proprietary machine learning and natural language processing model learns from hundreds of thousands of conversations taking place daily across their clients’ bots. This now allows Ada to consistently improve the accuracy and capabilities of its AI, and to ultimately strengthen the customer experience.
Chatbots can also be built without Artificial Intelligence. The hard-coded rule-based approach is technically easier, but lacks the scalability, flexibility and intelligence of the AI-based solutions. The caveats can be substituted by suggestive questions, combined with quick-reply buttons and other components, aiding the conversation with a less intuitive bot.
Machine learning vs. Rule-based Chatbots – Learn more about the differences between bots based on machine learning and rule-based bots here.
How to set Objectives for Your Bot ServiceWhen you have decided to begin a bot-building journey, the first step is to define the purpose, objectives and use-cases for the bot. In the customer care context, they may be used to accomplish customer requests – such as helping subscribers to configure the Access Point Name (APN), Wi-Fi hotspot or an email account, instruct them on how to top up their prepaid balance and so on.
Begin with Data ResearchResearch is the easiest way to assure that your chatbot will offer customers great value. Find out what kind of handset and service related problems your subscribers often experience, what handsets they use, which problems take the most time to solve, which handset models cause the most care inquiries, which are the most frequent routine questions, and so on. Many Online Smartphone Support platforms and call center work-flow management systems can provide this kind of data.
In telecoms, the problems that subscribers encounter are often handset dependent – whether it is about complex usability, a software bug, or a lack of knowledge by the user. Hence the conversation logic and help content on the bot must be optimized on a per device model or operating system basis. This is the compelling reason why it is so important to research the device population pattern on your mobile network.
One way to obtain a reliable view on the devices on your network is to capture real-time device detection data, usually easily available on network elements such as Automatic Device Detection (ADD) or Equipment Identity Register (EIR), and to cross drive it with Device Intelligence Data.
Device Intelligence Data provides information such as the handset brand and model, amongst hundreds of other capability data points about each handset.
Many Online Smartphone Support platforms provide detailed data on what kind of problems and which device models that telecom subscribers need the most help with. This data helps operators and MVNO to plan and optimize a Smartphone Support chatbot.
To Create or Outsource Support Content for your Bot?A bot-based customer service requires help content, which the bot can deliver to users upon request. Help content must be intuitive, self-explanatory and relevant for it to work properly. The content can be based on text instructions, screenshot images, interactive animation, how-to guide videos, or a combination of several media formats. Solutions to many problem situations can vary from a device model to another, or between operating systems. That’s why the help content should be tailored on a per device model and operating system basis.
Here are examples of smartphone help content, which are readily available for operators to use to save time and costs in the bot deployment:
- How to download and install application on iPhone X?
- What can I do if I have forgotten my Google Account password?
- How to setup an Email account on my HTC One A9?
Chatbot Communication ChannelsChatbots leverage chat mediums like SMS text messaging, Skype, the operator’s or MVNO’s website chat windows and social messaging services on platforms such as Facebook Messenger, Twitter, Kik, Telegram to receive and respond to messages. The most frequently encountered problem at telecom call centers is a non-functional Internet connection, which makes SMS text messaging an important back-up communication channel for chatbots.
Mobile Network IntegrationsA bot is a seamlessly embedded component on a mobile operator’s end-to-end customer care flow. This requires integrations into several external systems.
Here are some examples and links to further reference information:
- Integrations to CRM and ticketing tools can enable a swift chatbot to live agent handoff. For example, Ada partners with Zendesk, Genesys, and Live Person.
- An Automatic Device Detection system (ADD) and Device Intelligence Data repository can provide vital handset related information to bots.
- Online Smartphone Support platforms provide ready-made help content tailored for different device models and operating systems.
- An SMS Center opens a back-up communication channel, which will work even without an internet connection.
- An Equipment Identity Register (EIR) provides bots with information about stolen devices.
- VoLTE Device Entitlement Server (DES) gives operators the ability to deliver users Voice-over LTE and Rich Communication Service (RCS) configurations via the bot.
How to promote the Smartphone Support Bot?When your bot service is finally deployed and online, how can you get subscribers to use it, and not to contact the call center whenever a problem rises? As with any self-care service, customer engagement and promotion is crucial for driving up the usage, and you will eventually start cashing in the cost benefits.
In this area, mobile operators have an advantage as the chatbot service is included in the customer onboarding process
A definitive implementation alternative is that any operator can easily build is to combine the Smartphone Support chatbot with the automated device configuration process. Whenever a device is provided with the service settings package over-the-air, a SMS message is submitted and provide a direct URL link to the bot.
This introduces the bot service to subscribers right after the device configuration process, which often is the moment when device setup problems appear.
An automated device configuration process detects devices that are joining a network in real-time and delivers the correct service settings over-the-air. By integrating the Smartphone Support chatbot on that process, subscribers are delivered a Self-Care engagement message automatically after each device configuration. The SMS message contains an embedded URL link to the chatbot home page. This enables subscribers to solve any problem that may have occurred during the setup, or to get help with further device configuration.
ConclusionsChatbots have proven their benefits in the first line in telecom customer service; they save customers’ time, decrease support costs, and enable agents to deliver more meaningful human-to-human experiences. Now, first line customer service bots can be the beginning of a user-friendly transformation towards a fully automated telecom customer service!
This article aims to help operators and MVNOs on their bot building journey. It is updated at frequent intervals with new information – Stay tuned for new tips and feel free to contribute!