A chatbot, short for chat robot is, put simply, a computer program that is designed to stimulate a conversation with a user over the internet.

The chatbot software application processes text presented by a user and responds according to a set of rules or specific commands through to a series of routing blocks, for the purposes of simulating a conversation with them. Algorithms used can interpret what the user has said and determine an appropriate response.

Chatbots are used for entertainment, customer service and of course sales and marketing, where they give the user a more convenient way to access information without having to spend time trawling through content.

Chatbots are not about fooling the user into thinking they are talking to a real person, but more to engage with customers and make the interaction between a business and users more convenient and even have a bit of fun!

Is Chatbot Technology The Future Of Customer Support?

Across the world, chatbot technology means we have the capability to order a pizza, a taxi, book concert tickets or buy a new phone, all through our messenging apps such as Facebook Messenger.

Chatbots are quickly becoming the future of customer support, however we are still in the early days of the technology. So, while there are many easy-to-use platforms out there to simplify creating chatbots, there is still a lot of growing pains to work through to ensure a seamless integration.

Different Types Of Chatbots

Platforms such as Chatfuel & ManyChat allow us to build automated chatbots which don’t need any coding, and these are known as “Sequential Chatbots” or “Rule-Based Chatbots”.

Also known as “dumb bots”, these chatbots follow a flow or rules defined by the maker of the chatbot. The bot cannot go out of this scope. These chatbots result simply in a “Question and Answer” session.

Sequential bots are useful for capturing leads and feedback and answering simple FAQ’s. To get a better idea of this, think about the times you have spoken to Siri. She can pick up most things but also can cause frustration when she doesn’t understand what you’re trying to ask.

Chatfuel & ManyChat are ideal for building simple bots and do not suit complex commercial purposes.

When it comes to choosing which platform to build your bot, if you’re looking for a simple sequential bot then many of these platforms are very similar and it’s worth using the free versions to get a better idea of how to use and choose your preference from there.

Ideally, personalised programming is needed if you would like to use a chatbot to replace your customer service team or to deal with complex requests.

These specially programmed chatbots use machine learning and artificial intelligence which understand language and not just commands, getting smarter as it listens to conversations, and these are called “Intelligent chatbots”.

These bots provide the opportunity to users to ask any questions they have. If trained to answer that question, it will. A chatbot is trained by using data, machine learning and previous interactions.

Intelligent chatbots contain AI technology and Natural Language Processing (NLP) algorithms, which can understand what the user is trying to say and understand the intent. The bot is smart enough to understand the spelling mistakes of the users and understand the actual question of the user if poorly written.

Enabling languages other than English can also be done using AI & Machine Learning, although it is a challenge.

Example of a sequential chatbot:

West Coast Pet Care Centre Chatbot

Advantages Of Chatbots:

  • Allow conversation to happen outside of your normal trading hours – chatbots can be asked questions 24/7.
  • Chatbots never get tired, never need a break and continue to obey commands.
  • Chatbots help automate tasks which are to be done frequently and at specific times.
  • They can simultaneously have conversations instantly with thousands of people.
  • A customer support chatbot in your business to cater to simple queries of customers that then pass complex queries to customer support executives can help reduce staffing costs
  • You can communicate with your website visitors –  set up a bot to emulate live chat and talk with your customers.
  • They always treat a customer in a polite way.
  • They are fairly easy to make for developers, thanks to the presence of the various bot development platforms.
  • You can start building a chat bot directly through the Facebook Messenger’s development page.
  • The cost of making chatbots is not very high.

Disadvantages Of Chatbots:

  • Lack of security and data handling – it is quite easy for hackers to create a bot pretending to be a legitimate company, which can then convince unsuspecting users to provide personal information that can be used against them in phishing attacks or other malicious ways. For example, many instances of a “female” malware bot have been found on the Tinder app. Another example of a social engineering attack was the ‘Hitler was right’ PR disaster. That was what Microsoft Tay started to tweet within a day of its launch in March 2016.
  • Chatbots can only handle queries and comments from humans as long as the overall conversation flows in the ‘expected path’. When something else gets asked and the bot cannot respond in the expected way, the user becomes irritated.
  • Could cause possible confusion, if the user asks for specific products or information but are given something different; this may affect their purchasing decision.
  • At this time, there is still no chatbot that has passed the Turing test (a test for intelligence in a computer, requiring that a human being should be unable to distinguish the machine from another human being by using the replies to questions put to both).
  • Chatbots brings the risk of conversations being frustrating for potential customers and is inferior to what a human conversation would be.

Quick Guide On How To Build The Best Sequential Chatbot

  1. Start with a brainstorm and map out the questions you think your users would ask. Also map out the answers that you already have (eg from your FAQ page) and then jot down possible questions that could be attributed to those answers.
  2. Identify the intents, entities and responses of all.
  3. Learn about your users’ behaviour with your first chatbot so that you can tweak future versions. Start small!
  4. Think about creating a personality for your bot. How should it address your users? Does it say hi, what’s up, or how do you do? Does it have a name? (it should!) Even come up with a profile picture to go with it’s new name and personality. The more fun you make it to interact with, the more users are likely to forgive it if it can’t handle requests properly..
  5. Don’t try to pretend your chatbot is a real person. Users will see through that. Give the bot human-like qualities, and you can even joke about the fact it’s a bot.
  6. Compose 2 or 3 different greeting messages that explain the scope of your chatbot’s capabilities to set visitor expectations.
  7. Offer your users a simple selection of choices to help both your bot and your user.
  8. Personalise the messages with the user’s name.
  9. Throughout the conversation, ensure that your visitor never has to guess what to do next to get the information they need. Allow the customer to pick the next step.
  10. Minimise the risk of problems by preparing chat bots to introduce humans into the conversation if required. EG: “Let me connect you with our Customer Relations Executive”.
  11. Make sure your plan includes a test! Ask anyone you can to interact with your chatbot before it goes live. Collect comments and failed responses to finetune your keywords and phrases.

Example of a Weather Chatbot

Chatbot Glossary Of Terms That Marketers Should Know

When it comes to chatbot lingo, there are likely to be some unfamiliar terms. If you can only get your head around a few, these three are the main terms you need to know to become a chatbot marketer.

Utterance

This is what the user (human types) e.g. “what is the AFL score now”.

Intent

This is the reasoning behind the utterance – why did the user type that in the first place? In our example, they are engaging with the chatbot because they want to know the AFL score.

Entity

There are countless ways that humans can ask for the same data, which is where entities come in. These are the main words that the chatbot detects in order to understand the intent behind the utterance, cutting out any unnecessary or filler words. Whether somebody types “what is the AFL score now” or “show me the score in the AFL now”, the entities “score”, “afl”, and “now” are pulled in, and the chatbot returns the same results. At times, an entity can also be called a ‘slot’.

 

Additional “good to know” terms are as follows.

 

Broadcast

As opposed to a direct response to an utterance, a broadcast is exactly what it sounds like – a bulk message sent out to all subscribers, usually promotional.

Channel

These are the different places where a user can have interactions with a chatbot. The more popular mediums include Facebook Messenger and online websites (usually in the little pop-ups at the bottom of the page).

Conversational UI (User Interfaces)

As opposed to chatbot responses which look very much like an automated response, such as one that include buttons or emojis, conversational UIs appear are more in line with how a normal human would write or speak.

Natural Language Processing (NLP)

This is a process where computer software can understand a human’s text. For chatbots, NLP looks at an utterance, and from this, pulls in the intent and entities to understand how to respond.

Pilot

This is the testing phase before a chatbot gets released to the public, where a small number of people are engaged to look for any flaws or areas of improvement.

Response

The reply of the bot in relation to the utterance.

For Example: What’s the weather in Perth tomorrow?

Intent → Weather check

Entities → Perth (Location), tomorrow (Date)

Response → “The weather in {Location} {Date} is so and so”

What Else Can A Chatbot Do Other Than Answer Questions?

Chatbots are a useful tool for marketers due to the fact they can be used to automate scheduled messages to a database.

We can do this straight from the platform which connects through to Facebook Messenger. Facebook allow scheduled bots for the following:

  1. News: To inform people about recent or important events or information.
  2. Productivity: Enable to manage users productivity such as managing a calendar and reminders such as paying a bill.
  3. Personal trackers: Allows users to monitor information about themselves such as in fitness, health and finance.

Users must “opt in” to the messages by interacting with a bot in the first place. That’s why you may see the words “Get Started” when you open up a message to a company for the first time.

There are a number of different ways that you can broadcast messages:

  • To subscribers of a particular sequence.
  • To users who’ve been tagged in a past message.
  • By utilising demographic targeting within your user base.

Settings include:

  • Schedule Broadcast: Send an immediate message, or schedule it for a specific time and date.
  • Timezone Settings: Determine whether all subscribers receive the message at once, regardless of their time zone, or if they receive it at a specific time local to them.
  • Notification Settings: “Silent Push” notifications only show the phone notification, while a “Regular Push” notification comes with an accompanying sound notification.

Example Of How You Can Use A Chatbot

Marketers can use chatbots in a variety of ways, and ensuring they have a database of people to communicate with is key, which means you need to entice them to interact with your messenger bot. An example of this could be:

  • An ad on Facebook leads prospects to a key landing page on your site promoting a webinar.
  • When they fill in the form to register, the confirmation message also comes with a pop-up which asks them “Want to be reminded in Facebook Messenger?”
  • Upon clicking the pop-up, the Facebook Messenger app opens, and a bot asks them to click the “Get Started” button.
  • By clicking this button, the user is agreeing to their subscription with the bot, which can interact with them through automated messages and reminders as programmed. Note that the user should be aware that they will be engaging with your bot through their own Facebook profile.
  • After that, your initial welcome message appears.
  • You can then send or schedule Facebook ads through your chatbot platform directly into Messenger to all of your active users.
  • You can also engage new active users on current posts / videos via the comments.
  • A popular way of doing this is by inviting the users to comment a certain phrase on their post, like “give me a bonus book”. By doing this, the bot automatically sends the Get Started message.
  • Clicking Get Started means they have opted in!

Creating A Professional Chatbot

Messaging apps are overtaking social media, so you shouldn’t shy away from chatbots due to their complexity.

Test your bot in the real world and remember that if it continuously fails to appropriately respond to users’ requests, it can cause more damage than good to your brand. When it comes to customer service mistakes, most people are quite unforgiving, so you need to be positive that your chatbot works to improve the user’s experience, not hinder it.

Above all, ensure that your bot helps. If you help your potential customers, it will help your business!

As mentioned, communicating with your customers and users needs to be treated delicately, as not to hinder any potential purchasing decision.

Your chatbot is a voice and image for your business and you need to ensure it can deal with your user’s expectations while delighting them at the same time.

So, if you’re thinking about building a chat bot for your business, ensure you enlist a strategic marketing team like us to create it for you! You can view Living Online’s full range of services here.

PS: If you want more handy marketing tips, why not sign up to our newsletter? Just submit your email below!