
The Chatbot Beginner’s Guide: All Your Questions Answered

If you browse the internet, there’s a good chance you’ve come into contact with a chatbot. These conversational programs have proved a popular application of advanced tech, such as machine learning and natural language processing (NLP). And all the signs are there that chatbots will continue to play a role in business for years to come – so learning as much as possible about them means you’re well-placed to benefit in the long-term.
In this essential chatbot beginner's guide, you’ll learn:
Chatbots are computer programs with a persona – that of a robot (often a square-headed one with antennas). These robots’ primary purpose is to communicate with humans via text, voice, and touch.
Chatbots have been popping up all over the place for years now – literally, since you’re most likely to encounter them appearing at the bottom of webpages asking if you need any help.
Other possible places to meet these polite programs are messaging platforms (e.g. Skype, FB Messenger) or native apps, most commonly meant for employees and customer service. In fact, by Facebook’s 2018 measurements, there are 300,000 active chatbots on Messenger alone.
Chatbots, no matter where you find them, are designed to have conversations with humans. These conversations range from the rudimentary (e.g. “Tell me what time the train arrives”) to the more advanced (e.g. the likes of Alexa and Siri).
Here are some things chatbots can do:
They’re able to perform these actions by using tech to first process and understand what the user is saying. Then, they call on their “training”, possibly a database, to match the appropriate answer to the user's question. More advanced chatbots are able to understand context and intent – and, learn more about their human interlocutors and their preferences over time.
There are basically two types of chatbots:
The first type is your basic question-answer, decision tree program, that can only hold a conversation as long as the user says the right thing – and this is as good as it gets with them. An example would be asking the chatbot to pull an article from a knowledge base or asking how much a particular product costs.
The second type uses more powerful artificial intelligence, machine learning and predictive analytics, and are therefore better equipped to “sound” human and learn as they go.
These bots analyze a user’s input and employ technology like natural language understanding (NLU) to determine what the user means instead of just what they said. Then, they’re able to generate their own appropriate responses (think natural language generation or NLG). Understanding context is perhaps the key function of advanced AIs.
Chatbots can help organizations smoothen their operations, engage customers, and handle requests and complaints faster. More specifically, they:
For instance, Chatbots Magazine predicts that chatbots have the potential to automate 30 percent of the tasks done by today’s contact center staff. Also, Juniper research indicates that, by 2022, chatbots will save companies about $8 billion per year in customer supporting costs.
But more importantly, chatbots help companies increase their responsiveness and personalization abilities which, in turn, leads to better customer engagement and satisfaction.
Chatbots aren’t always the “computer brains” you expect. They’re only as good as humans make them – and that means both developers and users.
For instance, there’s the well-known example of Microsoft’s AI, Tay, which was turned from friendly to bigotted by Twitter users. Its machine learning capabilities were its downfall, as Tay absorbed the “bad” data the audience fed it.
Another problem with chatbots is when they simply can’t understand what the user says. If there isn’t a proper escape clause (e.g. “would you like me to connect you to one of my human colleagues” or “here are some options to choose from”), then the chatbot will fall into a loop with the user, constantly saying something akin to “I don’t understand, could you rephrase your request”.
And of course, there are always privacy considerations with software that handles user data. Users may not trust a chatbot that will show it knows their address and preferences if they don’t know how that data was collected.
Chatbots should always be secure and transparent, while maintaining a user-orientated approach. For example, some bots ask you to fill in your name and email in order to talk to you. This may have to do with security provisions or data capture methods, but it doesn’t make sense to be the first thing a chatbot says (considering the user hasn’t even said what they want or given their consent to be tracked).
Chatbots are used in a variety of settings inside organizations. Theoretically, all routine issues (e.g. questions that have simple answers) can be automated through digital assistants. And, AI-powered chatbots can go even further than that, handling actual human conversations with customers.
Below, you can find common use cases for business functions and industry.
Support and Engagement | Sales and Marketing | Employee management |
Provide first-tier support | Make personalized recommendations | Assist in finding information fast |
Answer repetitive questions | Facilitate email collection and marketing campaigns | Serve as project support, e.g. sending reminders |
Help visitors navigate sites | Support front-line sales and lead generation | Work as HR assistants |
Offer fun conversations | Handle purchases and onboarding | Help with employee onboarding |
Collect customer feedback | Perform inventory checks and track orders | |
Give access to information |
See the ultimate chatbot guide on chatbot use cases.
Here are some chatbot use cases for the banking and finance industry:
Fun fact: The famous Bank of America chatbot “Erica”, which launched in 2018, has served more than 10 million users since then and was able to understand close to 500,000 question variations by mid-2019.
Here are some insurance chatbot use cases:
Fun fact: 74 percent of consumers are willing to receive computer-generated advice about the type of insurance to purchase, 78 percent about investment asset allocation, and 68 percent about retirement planning.
Here are some telecom chatbot use cases:
Fun fact: IBM found that 56 percent of telecom customers use self-service options to choose the best plan and 77 percent use self-service to pay bills or recharge accounts.
Here are some ecommerce and retail use cases:
Fun fact: companies like Canadian automotive group Dilawri and furniture retailer Dufresne have brought their showrooms online and powered their services with chatbots, video chat, and more.
Here are some education use cases:
Fun fact: By using the chatbot Pounce to send personalized reminders and walk students through school processes, Georgia State University was able to reduce summer melt by 19 percent in the first year of implementation.
Here are some chatbot use cases in healthcare:
Fun fact: The global healthcare chatbots market is expected to grow at a CAGR of 25.1 percent from 2019 to reach $703.2 million by 2025, with “symptoms checking” accounting for the largest share of the market.
No matter how organizations use their chatbot, there are certain factors that must be considered. Namely, a successful chatbot should be:
Who says a bot can’t be informative and pleasant at the same time? From greetings to humor, a chatbot’s personality and conversational design can make it more attractive to users. Be sure your chatbot speaks in a natural way, though – as close to humans as possible without impersonating them (trying to pass off your bot as a human may end badly).
What happens when a user asks a question the chatbot doesn’t understand? Or, how should the chatbot move the conversation forward if the user doesn’t know what they want exactly? This is something companies need to take into account by adding contingencies, such as a direct link to humans, or sample questions the user can just click on. A chatbot with AI capabilities can be programmed to learn how to react properly on its own.
People want to interact with chatbots wherever and whenever it’s most convenient. Whether it’s Facebook Messenger or WhatsApp, Viber or Skype, your chatbot should be available to talk to your audience on your audience’s terms.
Failing to integrate chatbots with other systems is a missed opportunity. Sales software, support platforms (including your knowledge base and live chat), workforce management systems, data analytics, warehouse systems – and many other pieces of tech organizations use – can help chatbots do their job better by giving them access to a wealth of information.
Data security is governed by several local and international laws (e.g. GDPR). Chatbots need to be able to deal with privacy restrictions (e.g. by informing users of how their data is handled and asking them for consent). They should also have all necessary security provisions weaved into their code. Security-related questions to ask before buying a chatbot include:
Chatbots are a big part of many companies’ digital transformation efforts, so there’s no reason to assume your company will never need one. However, implementing a chatbot should come only after you have a clear idea of the challenges it’ll help you resolve.
In general, a chatbot is a good idea when:
Note: This list isn’t exhaustive; there are many ways you can use a chatbot to get real business value.
Chatbot design isn't a simple process – it requires planning and a lot of testing. But, it can be broken down into four generic steps:
What will your chatbot be doing? Will it provide support to customers, help them place orders, or ensure they can navigate your website? Perhaps all of the above? Be clear about your chatbot’s field of expertise. Also, give it a name and personality (will it be occasionally cheeky or always serious?).
Creating conversation flows (or dialogue trees) between users and chatbots is essential, especially for rule-based chatbots. Think about the end goal of the user – for instance, booking a hotel room. Now, how would the user get there? They could start with asking for prices during the month of May, or perhaps about room availability. Then, one user might continue with booking directly, while the other might request hotel reviews.
The most common user journeys need to be accounted for. Start with the ‘trigger’ event: what question/request/button click will trigger a particular sequence and chatbot greeting? Then, continue building the outlines of the dialogues. For example, from “Question about pricing”, you could go to “Book a demo” or “See feature list”.
Also, the bot's answers could differ depending on what has come before. This might result in complex diagrams, so try to start small.
You can use software like Draw.io, Google Drawings, and Microsoft Visio to help you build sequences.
Anticipating what the user could say in specific situations is important to address their needs. Know which dialogues to expect. Take the flows you’ve collected and flesh out the type of requests or remarks the chatbot might handle. The key here is to also consider alternative ways of phrasing utterances. Different users may use different words. While you should think about the concept of the question, you should also think about the possible sentences that go with it.
Create your lists of situations-questions with alternative phrasing (you could use a spreadsheet to start). Also, consider abbreviations, slang or commonly misspelled words if possible.
First, pay attention to the chatbot’s greetings. First impressions count with bots, too, and you want to make sure you start off on the right foot with users. While building greetings, stay on brand and use friendliness to draw the user in. Same goes for goodbyes – leave the user on a positive note and with an invitation to come back.
Then, you need to craft the responses to the questions you’ve identified looking at the flows and additional questions that have come up. Alternative questions will often have the same response, so the response should cover multiple phrasings. If that’s not possible, craft the individual responses for each of the alternatives.
Whatever you write, it’s good to keep it short, be direct, and use humor only when appropriate. Another thing to look at is predefined inputs or “quick replies.” Bots will usually let you craft possible responses to present to users. So, instead of typing their own sentences, users can click on buttons and trigger specific sequences.
More on writing chatbot scripts.
Chatbots have different levels of adoption among different industries. One thing is for sure, though; they’re here to stay and evolve. The hype might start to decline, but that doesn’t mean AI assistants will disappear – on the contrary, they’ll become a norm in most industries, especially in customer service.
And, this means two things: one, your company should jump on the chatbot train if you haven’t already, and two, creating a chatbot that works well now is a definite competitive advantage.
That’s where platforms like Acquire come in. Acquire is a robust customer engagement software that, among its other uses, lets you develop sales and support chatbots for your site. The steps we mention above, like building scripts and questions, can happen inside Acquire – and you can also turn on the ability to learn from interactions with users, as well as integrate your chatbot with AI services from IBM, Microsoft, and more.
Learn how Acquire improves support and
increases sales