The Definitive Guide to Creating a Successful ChatBot Strategy
If you – like me – spend time brushing up on the latest trends in customer service and support, you’ve likely heard a lot of chatter about chatbots. Much of the hype about chatbots is, in fact, warranted as they are providing companies a way to better interact with customers throughout the buying journey and through post-purchase. The benefits are real, but getting started and developing a strategy for them can feel daunting. This guide will help you sift through the features you really need, how to analyze costs and returns, determine the best platform for implementation, to ensure your chatbot initiative is a successful one.
What Functionality Do You Need?
The functionality you need from your chatbot largely hinges on the type of business you operate, or in some cases, the specific department the bot will serve. A retailer, for example, will likely benefit substantially from a bot designed to help customers purchase products, while a software company may benefit from a bot facilitating customer support services. Depending on your goals, you might want a chatbot capable of:
- Scheduling appointments or meetings
- Placing orders and recommending products
- Responding to customer inquiries
- Providing updates about service outages
- Issuing notifications on news, weather alerts, or other important information
Businesses developing a chatbot strategy should ask several key questions including:
- Are you aiming to facilitate in-house communication or customer-facing interactions?
- How will a chatbot impact the customer interaction?
- How will it influence the overall customer experience?
- What functions will it serve?
- What existing processes can be replaced or streamlined (without negative impact on customer interactions or experience)?
- What existing processes can be replaced?
- What resources will a chatbot require? Where will it get information? What other services and data does it need to interact with in order to ensure seamless access to the up-to-date information for users?
The answers to these questions will help you determine what functionality your chatbot requires tomeet business goals. For instance, if your chatbot strategy involves a bot answering customer questions, it will require access to information databases such as your existing knowledge base, help center, or FAQ.
Evaluating Costs and Returns
Chatbots vary widely in terms of costs, depending on the platform you use to build your bot, the amount of ongoing support you may need, and the number of messenger services where the bot is deployed. Pricing may also be based on the number of interactions. Chatbot technology has been around for decades, though the early iterations were quite robotic and lacked the artificial intelligence (AI) capabilities brands take advantage of today.
Still, because the core technology is well-established, it’s already possible for individuals to create their own personal chatbots serving as everything from a personal assistant to a financial data analyst/advisor like Digit, or even a friendly companion, starting at around $25 monthly. These bots, of course, are far more basic than the advanced, intelligent bots required to power customer support chat for a large enterprise.
Like any business investment, weighing the costs against the benefits and possible returns will drive the smartest decision-making. In the recruitment field, for example, bots have been used for over a decade, offering customized responses to inquiries based on context. While a chatbot capable of providing contextual responses in multiple languages, 24/7, communicating with applicants on any device, recruitment firms realize a number of valuable benefits in return. Most importantly, chatbots reduce the need to have recruiting agents man the messaging system 24/7 – plus they often provide faster and more accurate responses, reducing the time it takes for applicants to make decisions.
Costs are also dependent on the number of features bundled in the overall package, which may include:
- Conversation logs – complete history of actual conversations
- Customizable appearance/avatars – customizing the look and feel of the chat interface itself as well as the avatar or bot identity
- Detailed reports and analytics – insights on outcomes from conversations, conversions, and other data
- Multiple platform availability – ability to run chatbots on Facebook messenger, Slack, your website, and other platforms
- Knowledge Management System (KMS) – enabling team members to manage bots
- Missing knowledge reports or feedback – letting you know when your bot is unable to answer questions or provide accurate responses due to lack of accessible knowledge
- Database access – ability to instantly access and retrieve real-time data on shipping, product availability, and other evolving knowledge
- Other custom requirements
It’s not the development of the bot itself that dictates the true cost – the basic code and user interface for a basic bot is readily available. It’s the other specifications that increase the cost of implementation and ongoing management. It’s these factors, though, that give your bot real usefulness like conducting transactions, carrying out actions, and streamlining workflows.
In the simplest terms, think of advanced bots as capable in some way of conducting two-way interactions rather than a one-way information feed. This is the functionality that will offer a real return on investment. To project potential ROI, tie your business objectives to the necessary functionality to estimate the cost of implementation and ongoing management. Tie those same business objectives to the benefits or KPIs you’re aiming for, such as a reduction in customer support staff or increased sales. Many businesses find that by narrowing chatbot functionality to their true business requirements, it’s possible to realize substantial ROI from chatbot implementation.
How to Choose a ChatBot Platform: Evaluating Tools, Messaging Services, and Other Bot Essentials
Like cost, the right chatbot platform hinges on your business requirements. You’ll need a platform capable of delivering on your needs and able scale with your business as well. There are multiple platforms for developers to build their own bots, like Wit.ai, Motion.ai, or Chatfuel, as well as platforms offering chatbot integration as a component of self-service software or customer support solutions.
Some bot-building tools are designed to create platform-specific bots. Facebook Messenger Platform, obviously, is meant to aid developers in creating bots for Facebook Messenger, which many companies are already utilizing. Others are aimed at building bots for Slack or other services like Telegram. Developers who want to take the from-scratch approach can turn to open-source resources like Pandorabot’s Rosie, an Artificial Intelligence Markup Language (AIML) framework. Some companies will have more confidence in a bot developed in-house, while others will turn to established platforms for ease of implementation.
If you require a bot to run solely on Facebook Messenger, for instance, one platform may meet your needs today – but what happens when your customers want to interact with your business on Slack? While it’s fine to work with a platform that doesn’t yet have all the capabilities you might require in the future, you should examine the platform’s history, growth patterns, and pace of adoption in terms of keeping up with the latest and emerging trends in the space. Your choice of messaging platform can make all the difference in reach and engagement, which are key to producing ROI.
Security, of course, is always a top concern for companies implementing new technology, particularly when that technology is meant to function on third-party platforms where you don’t always have the same level of control over your data. But chatbots pose a unique challenge when it comes to security, as evidenced by the now-infamous Microsoft chatbot snafu. While that particular incident was influenced by a third-party outsider, the fear exists that bots could eventually become so intelligent that they could go rogue – even deactivating their own kill switches. This is a potential concern Google is tackling, with research in the works to develop a panic button that capable of killing a rogue AI agent.
While it’s pretty unlikely that your chatbot will literally develop a mind of its own and go rogue on you today, it’s a viable concern for the future and one that should make any business carefully evaluate potential platform partners. Choose a platform that’s up-to-speed on security issues and has systems and procedures in place to mitigate any potential risks.
When you carefully devise a chatbot strategy based on a thorough evaluation of your business requirements and map those needs to chatbot functionality and the right KPIs that indicate whether your bot is streamlining the proper target processes, you can reap the benefits of ROI driven by innovation while meeting your customers where they want to engage with you: in the messaging services they’re using every day.